As an interviewer, I don't ask questions like these.
By a very large margin, for most of the time you are actively developing as a full time employee at a large tech company, you're working on integrating systems, either through a build language or extending your software to use an already existing API.
What are the most useful skills for these tasks? Being able to communicate effectively, feeling comfortable asking questions, feeling comfortable admitting when you don't understand something, and being kind and friendly with those you interact with.
Therefore, when interviewing new hires, I do not ask many technical questions beyond a general competency question.
A company full of people great at communication but too little technical expertise will fail. I think it takes both types - technical gurus who can generate useful information and solve the harder problems, plus connectors/communicators who gather information.
I get the feeling at many companies now, there are fewer truly interesting technical problems than there are engineers enthusiastic or desperate to exercise their hard-won CS knowledge. Most real-world work of commercial value requires skill of course, and experience, but is not especially algorithmic. I think the surge of interest in DS is that here people see an opportunity to do something mathematical day-to-day (but 10% or less of DS is the "sexy" stuff).
Much of the work I do that could be considered "data science" involves trying to write generic visualization tools that allows us to plot various datasets and the results from models we've built (a lot of 2D/3D/4D weather/atmospheric data).
>What are the most useful skills for these tasks? Being able to communicate effectively, feeling comfortable asking questions, feeling comfortable admitting when you don't understand something, and being kind and friendly with those you interact with.
I've never liked this idea or thought it was true. At the end of the day it doesn't matter how could of communicators people are if they don't have the skill or talent.
Give me a genius programmer on the spectrum over someone who smooth-talked their way through the interview but doesn't know shit any day.
What about the reverse? Genius programmer that's remote and can't speak the same language as you or your team?
I think either way, a completely non-technical "good communicator" or a "genius programmer" non-communicator make it difficult to get these tasks done.
Can't say I could answer all these questions, but I think they're really great and motivate me to spend more time studying statistics. With the rise in data science and machine learning I've noticed a lot of resources devoted to teaching people how to create and train models for a given problem, but less resources on how to interpret those models and on statistical inference in general.
It's quite rare to find folks in the industry with formal education in statistics. The folks that have it can run circles around everyone else, though. Data Science and Analytics and Machine Learning all use statistical theory when boiled down.
When I ran a Data Science team we primarily hired Physicists and Mathematics majors and not CS graduates for this reason. It mostly worked as intended (untapped source of excellent candidates), but some of them could not for the life of them pick up software development / writing manageable code with a team. They were so used to writing unmanageable scripts that didn't have long-term time horizons (think: horrors of hundreds and thousands of lines of R-stats code in a single file, uncommented, and brutal to look at, with the expectation we'll just plug it into R-script on the command line) that occasionally this became a very tough habit to break.
Still, besides that issue - which all novice programmers have, but Data Scientists come in at a much more senior level - it was a great market advantage, one that I think is largely gone today, as companies understand the need for statisticians and economists who really and truly understand modeling and math.
Physicists and Mathematics majors can only be better than a CS grad, if they come prepped with Softwrae Development experience. A CS grad can communicate better with the code they write and most of the time assuming it's straight CS have knowledge and experience of ML that exceeds a Physics/Math major.
Anecdotally, I've worked in places where this is a major source of contention - The Data Scientists treated as talented individuals (who produce broken solutions which work for cherry picked data sets), and anyone else is just a monkey who maintains and fixes the broken code.
learn AI, start a company with a bunch of buddies with the sole intent of being acquired for your talents.
There was a company that did this a while back, I don't even think they actually had a product, they just knew that one of the big four would pay millions to acquihire them.
I find it curious that machine learning is a branch of computer science given how much of it is actually statistical learning theory.
The go-to graduate level machine learning text "The Elements of Statistical Learning" was written by 2 statistics professors (one, Prof. Hastie, a fellow South African! :)
Granted, neural networks are often taught as a bolt-on lecture in statistics course machine learning modules but topics like regularization and a rigorous study of overfitting were born in stats departments.
I'd say the opposite. It's quite rare to see employees without a formal education in statistics.
That being said, there are few of them and they congregate to the big tech and finance names. It can look like everyone or noone has a formal education, depending on where you personally work.
I appreciate the insight into the types of questions that might be asked. Sadly, I would have to look the interviewer in the eyes and say "I can program good"
These questions are largely unanswerable in their current form. Like"estimate the probably of disease in a city given Nationwide has a low probability". Like, what the heck does that even mean. I could imagine a dozen answers. I can only guess they are trying to get you to think about Bayes probability an event occurs given X independent identical trial). But that's ludicrous, because on what world has anyone proved disease occurs uniformly at random?
The point isn't to answer the question correctly. It's to show how you think through the problem. There is often a "right" way to approach the problem and a "right" way to answer the question, even if there is no right answer per se.
How I so desperately wish that were true in the general case. Sadly with many large corporations there is one single answer they're looking for and damn the false negatives.
The point isn't to answer the question correctly. It's to show how you think through the problem.
That's what people always say. Like, as if they actually tell from 1 or 2 of these goofy whiteboarding sessions.
It seems the most they can actually glean from these sessions is, "Huh, looks like their problem-solving approach is different from mine. Plus, they come in not having the advantage of having the solution already ready-baked and laid-out for them, in whatever interviewing guide it was I found this problem. Doesn't that suck for them? Not to mention being under pressure from doing way too many, generally carelessly orchestrated interview tracks (in their non-existent spare time) while being stuck in some job they're anxious to quit. And getting distracted by me diddling with my cell phone while they're doing headstands for me. Gosh, why is it so hard to find decent candidates, these days?"
There is a lot of cynicism about these types of questions in this thread, but I suspect a lot of it is coming from programmers and not data scientists. As a DS candidate I always find these kinds of interview questions a valuable learning experience and a positive challenge. It also provides me some insight into the kinds of problems they're interested in, and the kinds of approaches they use.
When I mean "right approach", I mean that there are several possible approaches, but only one of them is doable in a few hours in the afternoons, where as the others might take a research team and several weeks. Being able to recognize the right approach, and to ask the right clarifying questions, is important for data scientists, in the same way that being able to choose the correct algorithm or data structure is important for (certain types of) software engineers.
Agreed - these questions are definitely more appropriate for research oriented positions. And actually, way more interesting in their own right than most programmer interview questions.
What rubs me the wrong way is when, applying for more or less "standard dev" jobs, one gets asked research-grade questions (i.e. "Solve this cute problem I was fed the solution to recently, either in grad school or in some handout my manager gave me just now. But which was actually an open problem in the literature for 5-10 years before a reasonable solution came down the pipe.")
For a job that ends up being, "Congratulations on making it through our interview process! Now would you mind debugging this {{horribly written X}} that no on here wants to touch with a 10-foot pole? Everyone here has to do 9-12 months of shitwork before moving on to anything interesting here, you see."
I've never understood how there is a "right way to think about something". Isn't it an attribute of humanity that we all think and approach problems differently? Are companies really looking for one single type of individual to clone across all their engineering teams?
I believe it's not that there is a single right way to think about a problem, but there probably are wrong ways to think about the problem.
For example, going through all the issues with the question and implicit underlying assumptions, talking about what you'd need to give a good answer, or the characteristics a good answer may have might not be exactly what the interviewer expected or was looking for - but such an answer could demonstrate your competence in the field and be a good answer.
On the other hand, trying to make up an answer, or brazen your way through with jargon while hoping the interviewer didn't notice would likely be a bad answer.
It's not that these companies are looking for a single kind of thought. On the contrary, I believe they value different kinds of thought. Instead, they believe they can sort kinds of thought into desirable and not.
The reality is, unless you're directly communicating with someone who understands the problem/question, they'll be expecting the textbook definition as is or whatever stackoverflow response has.
I have mixed feelings concerning this methodology. At the same time I feel I'm a bit rust on the details of the field, I still believe best way to assess academic background is by sending a paper "offline" and asking the interviewee to explain it.
IMO all these quick-answer questions asses is how prepared is the interviewee in answering quick-answer questions.
Agreed. I have horrible short term memory but these questions are trivial as a "homework" type assignment. Plus, for my current data scientist role, the interview consisted of me bringing in a project I had worked on (actual code) and demonstrating how it worked. I much prefer that tactic. I think it gives you greater insight into the candidate.
I'm sorry if I'm ignorant of the platform, are the different posts the interview questions? Also what am I to make of the platform itself? Is this an app hosted on heroku I can access later on my phone? If that's the case that would be awesome, like a targeted stackoverflow.
If you obtained a degree in Computer Science and specialized in Machine Learning are you suppose to be able to answer these questions? What job specialization is this aimed for? Almost strikes me more as a statistical based interview.
a degree in computer science is really inefficient if you want to be a data scientist. imagine a 10 class CIS masters: with graduation restrictions you might be able to take 3 classes that directly relate to data science. good luck in your job interviews if you took them first, as you spent 20 hours a week on homework to fill requirements you will never use.
now take a statistics major, every class is relevant, and you can still take machine learning in your electives. win win.
I came to this conclusion after I noticed more of my classmates in the mba program (wharton) as data scientists than people in computer science who took machine learning. in fact, _all_ of the CIS majors in machine learning who really wanted to be a data scientist ended up as engineers.
so then I started doing a small search on linkedin, only looking at the big tech company data scientists. selection biases aside, out of 12 profiles: 5 statistics majors, 5 business, 1 biophysics, 1 IT major.
I have also done some looking into interview questions via glass door, and you get grilled on statistics questions. this matches my one interview with uber in 2016. I only got asked 2 ML questions: what is random about a random forest, and in KNN, what happens to bias & variance as K goes to 1
if you want to be a data scientist, you need to learn stats really well or getting past the interview process is going to be very difficult.
No idea why this was downvoted; I find this to be an accurate description. The difference between stats and ML is mostly one of terminology and perspective. I studied math in college and grad school, and there were several moments in Andrew Ng's online lectures where I thought "oh, I know this, but we didn't call it that, and I had no idea it was considered ML."
Do you mind elaborating?
I can't think of any methods that explicitly (or implicitly) use p-values.
The only place in the industry I have seen p-values used is AB-testing, but most seem to be trying to move to a multi-armed bandit, bayesian methodology.
Machine learning systems that perform a job and make money are still overwhelmingly stats-based, and the knowledge required to understand, tune, and optimize any ml system, regardless of design, are based on statistics and always will be.
I'm not sure I wholly agree there. There's an awful lot of real work being done that makes real money based on deep learning, and deep learning comes from the CS "it works, what's the problem" side of the field as opposed to the statistics "I want formally grounded theory for everything" side.
One of the big pushes in Bayesian statistics recently has been to try to figure what the hell all these neural nets are actually doing. It's certainly not the case that the stats have been in the driving seat there.
I have the impression that at big universities maths is always the #1 topic in CS/ML. So its no surprise their graduates ask the same riddles as their profs.
People have been doing statistics for decades, which is mostly all that machine learning is. Coming up with a fancy new buzz word for stuff that has been for decades, doesn't make something "elite".
You know, statistics is a branch of mathematics, and mathematical education itself can be considered "elite", if for no other reason then for the irrational fear it induces in most of people.
Most of high school graduates decides to pursue more "practical" topics like CS, leaving mathematics to those chosen few who dared to study it :)
Then why do machine learning scientists get paid multiples more than statisticians? Why do CS departments get far more funding for ML grants than the same grant branded from the statistics department?
Say what you want about ML vs statistics, the fact is: in both academia and industry, machine learning is far sexier than statistics.
Machine learning systems that perform a job and make money are still overwhelmingly stats-based, and the knowledge required to understand, tune, and optimize any ml system, regardless of design, are based on statistics and always will be.
These are good questions. A good quantitative researcher can answer most of them. They're also fairly open ended, allowing the candidate to show what he or she knows.
My quick and dirty answers are below. I'm thinking of moving jobs within the next year so I could use the practice. Can anyone do better?
> How do you treat colinearity?
Throw away the redundant part of the data
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
This is very problem-dependent, but it's got the potential to wreak havoc with your learning algorithms. You might get seemingly good results by e.g. always predicting positive. Think carefully about your loss function.
> How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives, profession and gender? Define a machine learning algorithm for this.
This is a messy combination of continuous, categorical, and binary data. I'd encode the data in a vector [salary, x-y co-ords, one-hot encoding of profession, and binary indicator for gender]. Something like a random forest will probably get you most of the way there. Unusual professions could mess with the algorithm, so consider grouping by industry or averaging with a model that omits the profession data.
> Is it useful to apply PCA to your data before SVM classification?
Probably, but it could be data dependent. Semi-supervised learning often improves machine learning models, and it'll lower the dimensionality of your inputs, which will make training / hyperparameter search far more efficient.
> How do you compare a neural network that has one layer, one input and output to a logistic regression model?
Not entirely sure I understand the question, but a NN is basically just nested logistic regression, depending on the activation function.
> From a long sorted list and a short 4 element sorted list, which algorithm will you use to search the long sorted list for 4 elements.
To be honest, I'd use binary search and call it a day. I realise this isn't the answer you're looking for.
> How will inspect missing data and when are they important for your analysis?
You might try to impute the missing data from similar datapoints. Maybe use something like K nearest neighbours on the non-missing components to deduce the missing values. Got to be careful doing this though, since it could massively bias your analysis: consider using a special encoding for "missing data" too.
> Estimate the probability of a disease in a particular city given that the probability of the disease on a national level is low.
This is a tough one, since there are so many ways to answer it. I think I'd start by writing down a list of factors that might allow a high prevalence of the disease locally despite a low prevalence nationally. Weather? Local wildlife? Proximity to major transport hubs? Then I might suggest a simple model like naive Bayes.
Use an LASSO as a feature selector or regularizer.
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
User F1 score instead of accuracy as a performance metric. Try using cascade classifier.
> How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives, profession and gender? Define a machine learning algorithm for this.
More or less, the same as what you said.
> Is it useful to apply PCA to your data before SVM classification?
No, since kernel-SVM optimization is dependent on data (similarity) and not features. Also, we will be projecting the data into a basis expanded space anyways, so the dimensional reduction would be redundant.
> How do you compare a neural network that has one layer, one input and output to a logistic regression model?
An MLP-NN is the same as a logistic regressor, if all connections as 1-1 (no basis expansion), no activation function in the data-1st layer connection and Sigmoid as activation function in the 1st layer-output connection.
> From a long sorted list and a short 4 element sorted list, which algorithm will you use to search the long sorted list for 4 elements.
I am dreadfully bad at core CS questions...throw some AI questions at me. XD
> How will inspect missing data and when are they important for your analysis?
You might try to impute the missing data from similar datapoints. Maybe use something like K nearest neighbours on the non-missing components to deduce the missing values. Got to be careful doing this though, since it could massively bias your analysis: consider using a special encoding for "missing data" too.
Love your solutions. I might also try using cascade classifiers.
> Estimate the probability of a disease in a particular city given that the probability of the disease on a national level is low.
tough question. Would probably involve me asking a lot of questions about what data I have to play with. Very interview specific.
I am not a data scientist, just a researcher in cog neuro. With multicollinearity, I'd first want to know whether these variables are meant to measure the same latent construct. If so, creating a latent factor in a SEM seems the way to go. Lassos and ridge regression I've heard of, but I know very little, except thatse methods can find use when using fmri voxel timeseries...
Regarding the PCA before SVM step - PCA would help reduce dimensionality w.r.t linear correlations, kernels generally are used to obtain non-linear maps. So wouldn't it be wrong to say doing PCA is redundant? Also, while kernel SVM optimization is dominated by number of data points, computing the value of the kernel function for individual data points might depend on the dimensionality itself, would it not? Another instance of lower dimensional representation possibly helping.
Please correct me if I'm wrong, I'm still just learning this stuff.
The cost of computing similarities is definitely linear or lower, wrt. to features. But overall, the Kernel SVM will scale to the order of n^3 wrt. to the number of points. In that sense, focussing on the number of features may not make a lot of sense.
PCA isn't lossless either. We are definitely trading off some richness in the data for efficiency. When the gain in efficiency is negligible (as is the case with SVMs) the trade off doesn't look as tempting.
I am student in the area as well, so take my words with appropriate caution.
2 things I would like to add to the already very good answers in this thread:
> How do you treat colinearity?
Look at feature interactions when using decision tree based models to discover if some interactions between seemingly similar features could improve the performance of the model.
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
Don't use accuracy or other "averaging" evaluation metrics. When looking at unbalanced data you are likely interested in the minority class so I would use metrics that only look at this class for performance evaluation.
Colinearity can be handled either through variable selection, or dimensionality reduction. The right approach depends on the data. If you're trying to identify the root cause of some phenomenon, L1 regularized models will pick out variables (though they might pick the wrong one). If you just want to prevent colinear variables from taking over your model, dimensionality reduction is the right choice.
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
This is one area where bayesian statistics really shines. The variance of the posterior predictive distribution for a bayesian model is inversely proportional to the amount of data you use for inference.
> How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives,
profession and gender? Define a machine learning algorithm for this.
Random forests are a good answer to this in terms of predictive accuracy. The problem with random forests is that they don't give you any information about how certain your estimates are. A better technique to use would be Gaussian processes (GPs). GPs are a kernel method somewhat similar to SVMs, but they have the marked advantage that they provide a posterior distribution over their estimates, rather than just the maximum likelihood value.
> Is it useful to apply PCA to your data before SVM classification?
The right answer is that it depends on the shape of your data and the specific SVM kernel/hyperparameters you're using. If you have a square matrix and you're doing linear SVM, PCA first will be slower. On the other hand, for a highly asymmetric matrix and nonlinear SVM, PCA done by stochastic approximation of the SVD, followed by SVM will be much faster. From a prediction accuracy standpoint it isn't going to make a large difference in either case since the kernel trick effectively projects your data into a high dimensionality feature space.
> From a long sorted list and a short 4 element sorted list, which algorithm will you use to search the long sorted list for 4 elements.
To answer this properly you need to know if the list has repeated elements. If there are no repeated elements then a binary search on the first element of the 4 item list, then scanning forward from there is the correct answer. If you admit the possibility of (possible massive) repetition of elements then this is no longer the case.
> How will inspect missing data and when are they important for your analysis?
This is yet another case where bayesian analysis shines (can you tell what I specialize in?). With bayesian methods you don't have to impute a value, rather you can treat unknowns as coming from a specific distribution (which can be the exactly distribution of that variable, via empirical bayes). This prevents you from introducing bias in your model.
> Estimate the probability of a disease in a particular city given that the probability of the disease on a national level is low.
This question is kind of bad given that it doesn't put a bound on what data is available to you. In terms of model, this is just a classification problem in disguise
I respectfully disagree with everyone saying these are statistics. This is analytics. Most of them are down to earth and practical matters that I think would be lost on someone who did statistics as a math degree (but maybe this distinction is unclear because it is also observably different from straight comp sci or neural net expertise).
I recently graduated from a masters of business analytics program and would have humbly assumed that the breadth and depth of knowledge asked for in these questions would be beyond my capacity, but it was actually pretty spot on with what I had learned. Strictly speaking, I am far from an expert in statistics/math/comp sci. But these seem like reasonable questions. The best one is
"How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives, profession and gender? Define a machine learning algorithm for this."
This is a difficult question because for someone who has played with data science before, the answer seems to be obvious: throw the data into your favorite predictive algorithm and see the result (worst case: deep learning). What they are probably looking to see is a discussion of feature engineering, hypothetical exploratory analysis, and an effective way to present the answer (straight probabilities, weighted loss functions, outlier detection, etc.)
It appears that the current generation of "hackers" seems very keen to work for a big name company-- Amazon, Apple, Microsoft, Google, Facebook. As someone who has worked for many startups and several of the above named companies, please let me give you a bit of advice:
- A big company on your resume has no outside value to your prospects.
Seriously. There's not "hey well he worked for amazon so we know he's good" free pass in the future for having Amazon on your resume.
There might be a "well, she wasted 3 years at amazon and we all know that's a hellhole so we can get away with abusing her" signal though. (Amazon is a hellhole, I spent 2 years there, in retrospect I would have been better off walking the day it became clear they were using cult methodology.[1])
It took me many years to find my calling. Once I did then my career direction was set. I'll only be working for companies in this industry going forward, because it really is my calling.
Take a job that maximizes your creativity and ability to contribute.
Don't chase big money or big names. The two almost never go together, anyway.
[1] Magical Phrases are a key indicator of cult methodology. Also known as "thought terminating cliches" they are multi-purpose sayings that are used to invoke company ideology and shut down dissent. At amazon there are things like "It's day one!" and the like. Every time I've seen these cult signals at a company it turned out to be a toxic work environment.
A good question to ask in interviews is if they company has any slogans all the employees know.
My experience is completely different from yours in regards to every single point you make. Working at a big company (amazon) has increased my salary significantly, made me better technically, and noticeably made it easier for me to find a job elsewhere afterwards (this could be because my technical skill were better afterwards, and I had more experience). Additionally, I actually loved my team at Amazon. My manager was incredibly empathetic and caring, far more than any managers I have had at smaller companies before. I just wanted to provide a different perspective, and show that generalizing about companies, especially large ones, is a bad idea and often grossly incorrect. Your experience and the people you talked to (likely) represent a very small part of the whole company.
I had a similar experience there. There were parts of the company that seemed like they sucked, but my experience was only positive and put me in a position to learn a ton from strong engineers while building my skills up and making a good amount of money. When I started, I made less than some of my friends who chose startups, but two years out of college I was making more than almost anybody I knew.
I might not have worked on my passion in specific there, but I had a great time, learned a ton, and IMO set myself up for a great trajectory in my career.
Like most other large companies, YMMV depending on where you're at.
First, because I think you're factually mistaken - for a lot of people, a big name on your resume is a huge, positive flag.
Second, because there are other effects of working at a large company which you're not considering. Most importantly - they're large! That means you're going to interact with a huge amount of people. In 3 years at a BigCo, you'll likely have contacts not only in various departments, but probably in several other companies as well, both because of work with outside companies, and because of interacting with employees who left.
Not to mention, you can learn a lot about the industry that you might not necessarily be exposed to while at a small startup - via interactions with those outside companies (customers, partners, etc). You can also work on seriously big things, which you won't be working on at a startup, for example.
None of this is to say that working at a BigCo is the right move. I actually tend to think the opposite, depending on various factors. But it's definitely not an all-downsides proposition.
> - A big company on your resume has no outside value to your prospects.
I respectfully disagree. Recruiters and hiring managers, and even individual interviewers take me far more seriously after having "Google" on my resume than they did before. It's not enough to land a good job automatically, but most interactions now at least start from the presumption that I'm competent, which makes a job search less frustrating.
I haven't worked for AppAmaGooBookSoft, but I know if we got an ex-Googler or similar at many of the places I've worked at, they'd be widely respected/feared by my peers on that virtue alone, and it'd certainly help a lot in the hiring process.
I know some well-known niche companies with highly-coveted positions that appear to make new hires almost exclusively from AppAmaGooBookSoft alumni (last time I mentioned this, a user correctly guessed one of the companies I was referring to from just such a vague description).
The reality is that most people hate hiring and have no idea whatsoever how to do it well (in the sense of bringing on competent people, not from the HR perspective of minimizing legal liability). Big names, in either education or employment history, help a lot. It's probably easy to underestimate how much they help if you haven't gone without them.
I would also venture that ex-BigNameCo is a lot less of a big deal if you work somewhere that's a hotspot for that type of company (Silicon Valley, Seattle). There it may be seen more as a sign of rejection than success, like "Why don't you still work at Google if you were Google material?". But even still, most companies are going to be swayed significantly by a big name on the resume.
> Don't chase big money or big names. The two almost never go together, anyway.
Do you have evidence on that? Virtually all data I've seen from salary sharing threads, and all compensation discussions I've had with startups suggests the contrary.
I disagree. I am tech lead at a small company, and when we look for help, permanent (rare) or contractor (usually), big software company experience is a definite plus. At least they have been vetted to some level by others first, and have probably had some decent exposure to current methodologies and practices. I really dislike looking for new help, and this experience provides a valuable data point.
Seriously. There's not "hey well he worked for amazon so we know he's good" free pass in the future for having Amazon on your resume.
The point of a CV is to get you an interview. A person who graduated from an Ivy/Russell, did a year or two at Google and a similar spell at Goldman Sachs, will be able to get an interview anywhere, for the rest of their lives.
They probably won't get an easier ride in the interview, it may even be harder as the interviewer's egos insist that they best one of the "elite", but they will never need to worry about passing a screen.
I used to be contacted by Amazon a lot around 4 years ago, to work with their data science team. At the time I was busy helping build up a startup (and a data science team), so always turned them down, I always put my people first. The startup turned out to be quite profitable, they are up and kicking nowadays, I am very proud of my time with them.
Nowadays I am looking for a new permanent position (have been freelancing for a couple of years). My wife happens to be a headhunter and knows how most HR and internal recruitment teams think and work. She tells me with Amazon in my CV many more doors would open. I don't regret turning them down, I know for sure had I accepted to work with them I would have never met my wife, who brings me more joy than any job would ever do... But I can't help to wonder...
People generally make more money at big companies especially when you factor in 401k and healthcare benefits. However impact is lower at big companies for the average employee. If you want a job with a large impact and good/great money then profitable medium-sized companies are where it is at.
As for name recognition, it always helps. No employer is going to hire someone solely based on where they went to school or where they worked previously. However a well-known school or company can be the deciding factor between two candidates who are otherwise equal. My wife went to Purdue and did her pharmacy residency in Colorado. She later found out that she was competing with three other candidates for one position who were equally as good in the eyes of the residency program, and she won because she graduated from a well-known, top 10 pharmacy school.
This doesn't really match up with my anectodal experience - the number of recruiter emails I get every week has more than tripled since I started working at Spotify (after working at a number of smaller startups before).
interesting.. I have worked both for startups and large companies. There are tradeoffs to both of course but my favorite so far is the large company I am at. I get to do things at a scale that doesn't exist at startups.
I think its already been iterated on by other commenters, but almost all of this advice is wrong. Name brands play a huge roll in careers, even in programming ones. If you have worked a long career and havent worked at a big name company, thats a huge negative signal.
As an interviewer, I don't ask questions like these.
By a very large margin, for most of the time you are actively developing as a full time employee at a large tech company, you're working on integrating systems, either through a build language or extending your software to use an already existing API.
What are the most useful skills for these tasks? Being able to communicate effectively, feeling comfortable asking questions, feeling comfortable admitting when you don't understand something, and being kind and friendly with those you interact with.
Therefore, when interviewing new hires, I do not ask many technical questions beyond a general competency question.
A company full of people great at communication but too little technical expertise will fail. I think it takes both types - technical gurus who can generate useful information and solve the harder problems, plus connectors/communicators who gather information.
I get the feeling at many companies now, there are fewer truly interesting technical problems than there are engineers enthusiastic or desperate to exercise their hard-won CS knowledge. Most real-world work of commercial value requires skill of course, and experience, but is not especially algorithmic. I think the surge of interest in DS is that here people see an opportunity to do something mathematical day-to-day (but 10% or less of DS is the "sexy" stuff).
Much of the work I do that could be considered "data science" involves trying to write generic visualization tools that allows us to plot various datasets and the results from models we've built (a lot of 2D/3D/4D weather/atmospheric data).
This is data science, not software engineering.
>What are the most useful skills for these tasks? Being able to communicate effectively, feeling comfortable asking questions, feeling comfortable admitting when you don't understand something, and being kind and friendly with those you interact with.
I've never liked this idea or thought it was true. At the end of the day it doesn't matter how could of communicators people are if they don't have the skill or talent.
Give me a genius programmer on the spectrum over someone who smooth-talked their way through the interview but doesn't know shit any day.
What about the reverse? Genius programmer that's remote and can't speak the same language as you or your team?
I think either way, a completely non-technical "good communicator" or a "genius programmer" non-communicator make it difficult to get these tasks done.
Can't say I could answer all these questions, but I think they're really great and motivate me to spend more time studying statistics. With the rise in data science and machine learning I've noticed a lot of resources devoted to teaching people how to create and train models for a given problem, but less resources on how to interpret those models and on statistical inference in general.
It's quite rare to find folks in the industry with formal education in statistics. The folks that have it can run circles around everyone else, though. Data Science and Analytics and Machine Learning all use statistical theory when boiled down.
When I ran a Data Science team we primarily hired Physicists and Mathematics majors and not CS graduates for this reason. It mostly worked as intended (untapped source of excellent candidates), but some of them could not for the life of them pick up software development / writing manageable code with a team. They were so used to writing unmanageable scripts that didn't have long-term time horizons (think: horrors of hundreds and thousands of lines of R-stats code in a single file, uncommented, and brutal to look at, with the expectation we'll just plug it into R-script on the command line) that occasionally this became a very tough habit to break.
Still, besides that issue - which all novice programmers have, but Data Scientists come in at a much more senior level - it was a great market advantage, one that I think is largely gone today, as companies understand the need for statisticians and economists who really and truly understand modeling and math.
Physicists and Mathematics majors can only be better than a CS grad, if they come prepped with Softwrae Development experience. A CS grad can communicate better with the code they write and most of the time assuming it's straight CS have knowledge and experience of ML that exceeds a Physics/Math major.
Anecdotally, I've worked in places where this is a major source of contention - The Data Scientists treated as talented individuals (who produce broken solutions which work for cherry picked data sets), and anyone else is just a monkey who maintains and fixes the broken code.
>Physicists and Mathematics majors can only be better than a CS grad, if they come prepped with Softwrae Development experience
This is very much the opposite of my experience.
Yeah, my experience is almost a perfect match. These days I'm wondering how a statistician can best capitalize long-term in tech?
learn AI, start a company with a bunch of buddies with the sole intent of being acquired for your talents.
There was a company that did this a while back, I don't even think they actually had a product, they just knew that one of the big four would pay millions to acquihire them.
I find it curious that machine learning is a branch of computer science given how much of it is actually statistical learning theory.
The go-to graduate level machine learning text "The Elements of Statistical Learning" was written by 2 statistics professors (one, Prof. Hastie, a fellow South African! :)
Granted, neural networks are often taught as a bolt-on lecture in statistics course machine learning modules but topics like regularization and a rigorous study of overfitting were born in stats departments.
I'd say the opposite. It's quite rare to see employees without a formal education in statistics.
That being said, there are few of them and they congregate to the big tech and finance names. It can look like everyone or noone has a formal education, depending on where you personally work.
I appreciate the insight into the types of questions that might be asked. Sadly, I would have to look the interviewer in the eyes and say "I can program good"
Data science now has its own tree-sorting questions. Shameless plug, but since the submission is popular, some could find this helpful: https://shapescience.xyz/blog/interview-questions-for-data-s...
These questions are largely unanswerable in their current form. Like"estimate the probably of disease in a city given Nationwide has a low probability". Like, what the heck does that even mean. I could imagine a dozen answers. I can only guess they are trying to get you to think about Bayes probability an event occurs given X independent identical trial). But that's ludicrous, because on what world has anyone proved disease occurs uniformly at random?
The point isn't to answer the question correctly. It's to show how you think through the problem. There is often a "right" way to approach the problem and a "right" way to answer the question, even if there is no right answer per se.
How I so desperately wish that were true in the general case. Sadly with many large corporations there is one single answer they're looking for and damn the false negatives.
The point isn't to answer the question correctly. It's to show how you think through the problem.
That's what people always say. Like, as if they actually tell from 1 or 2 of these goofy whiteboarding sessions.
It seems the most they can actually glean from these sessions is, "Huh, looks like their problem-solving approach is different from mine. Plus, they come in not having the advantage of having the solution already ready-baked and laid-out for them, in whatever interviewing guide it was I found this problem. Doesn't that suck for them? Not to mention being under pressure from doing way too many, generally carelessly orchestrated interview tracks (in their non-existent spare time) while being stuck in some job they're anxious to quit. And getting distracted by me diddling with my cell phone while they're doing headstands for me. Gosh, why is it so hard to find decent candidates, these days?"
There is a lot of cynicism about these types of questions in this thread, but I suspect a lot of it is coming from programmers and not data scientists. As a DS candidate I always find these kinds of interview questions a valuable learning experience and a positive challenge. It also provides me some insight into the kinds of problems they're interested in, and the kinds of approaches they use.
When I mean "right approach", I mean that there are several possible approaches, but only one of them is doable in a few hours in the afternoons, where as the others might take a research team and several weeks. Being able to recognize the right approach, and to ask the right clarifying questions, is important for data scientists, in the same way that being able to choose the correct algorithm or data structure is important for (certain types of) software engineers.
Agreed - these questions are definitely more appropriate for research oriented positions. And actually, way more interesting in their own right than most programmer interview questions.
What rubs me the wrong way is when, applying for more or less "standard dev" jobs, one gets asked research-grade questions (i.e. "Solve this cute problem I was fed the solution to recently, either in grad school or in some handout my manager gave me just now. But which was actually an open problem in the literature for 5-10 years before a reasonable solution came down the pipe.")
For a job that ends up being, "Congratulations on making it through our interview process! Now would you mind debugging this {{horribly written X}} that no on here wants to touch with a 10-foot pole? Everyone here has to do 9-12 months of shitwork before moving on to anything interesting here, you see."
I've never understood how there is a "right way to think about something". Isn't it an attribute of humanity that we all think and approach problems differently? Are companies really looking for one single type of individual to clone across all their engineering teams?
I believe it's not that there is a single right way to think about a problem, but there probably are wrong ways to think about the problem.
For example, going through all the issues with the question and implicit underlying assumptions, talking about what you'd need to give a good answer, or the characteristics a good answer may have might not be exactly what the interviewer expected or was looking for - but such an answer could demonstrate your competence in the field and be a good answer.
On the other hand, trying to make up an answer, or brazen your way through with jargon while hoping the interviewer didn't notice would likely be a bad answer.
It's not that these companies are looking for a single kind of thought. On the contrary, I believe they value different kinds of thought. Instead, they believe they can sort kinds of thought into desirable and not.
The reality is, unless you're directly communicating with someone who understands the problem/question, they'll be expecting the textbook definition as is or whatever stackoverflow response has.
In data science interviewing, that generally is the case. Technical people are going to be asking these questions and evaluating you.
This is exactly it. Thanks for laying it out clearly.
I have mixed feelings concerning this methodology. At the same time I feel I'm a bit rust on the details of the field, I still believe best way to assess academic background is by sending a paper "offline" and asking the interviewee to explain it. IMO all these quick-answer questions asses is how prepared is the interviewee in answering quick-answer questions.
Agreed. I have horrible short term memory but these questions are trivial as a "homework" type assignment. Plus, for my current data scientist role, the interview consisted of me bringing in a project I had worked on (actual code) and demonstrating how it worked. I much prefer that tactic. I think it gives you greater insight into the candidate.
bringing in a project I had worked on (actual code
I don't think commercial confidentiality makes that a viable generic method...
Should have clarirified. This was right out of grad school—so code for my research.
I'm sorry if I'm ignorant of the platform, are the different posts the interview questions? Also what am I to make of the platform itself? Is this an app hosted on heroku I can access later on my phone? If that's the case that would be awesome, like a targeted stackoverflow.
My interview was outrageously more math-y than this. Sadly, Amazon NDAs these things pretty hard.
If you obtained a degree in Computer Science and specialized in Machine Learning are you suppose to be able to answer these questions? What job specialization is this aimed for? Almost strikes me more as a statistical based interview.
a degree in computer science is really inefficient if you want to be a data scientist. imagine a 10 class CIS masters: with graduation restrictions you might be able to take 3 classes that directly relate to data science. good luck in your job interviews if you took them first, as you spent 20 hours a week on homework to fill requirements you will never use.
now take a statistics major, every class is relevant, and you can still take machine learning in your electives. win win.
I came to this conclusion after I noticed more of my classmates in the mba program (wharton) as data scientists than people in computer science who took machine learning. in fact, _all_ of the CIS majors in machine learning who really wanted to be a data scientist ended up as engineers.
so then I started doing a small search on linkedin, only looking at the big tech company data scientists. selection biases aside, out of 12 profiles: 5 statistics majors, 5 business, 1 biophysics, 1 IT major.
I have also done some looking into interview questions via glass door, and you get grilled on statistics questions. this matches my one interview with uber in 2016. I only got asked 2 ML questions: what is random about a random forest, and in KNN, what happens to bias & variance as K goes to 1
if you want to be a data scientist, you need to learn stats really well or getting past the interview process is going to be very difficult.
Statistics and machine learning have a huge amount of overlap. Almost seems silly we separate the fields.
Well yes. Before the trendy buzzword, machine learning was known simply as predictive statistics.
No idea why this was downvoted; I find this to be an accurate description. The difference between stats and ML is mostly one of terminology and perspective. I studied math in college and grad school, and there were several moments in Andrew Ng's online lectures where I thought "oh, I know this, but we didn't call it that, and I had no idea it was considered ML."
reminds me of the whole Bayesian perspective of ML
Or simply statistical modelling. I think there are methodological differences though. To me ML seems like a massive p-value fishing operation.
Do you mind elaborating? I can't think of any methods that explicitly (or implicitly) use p-values. The only place in the industry I have seen p-values used is AB-testing, but most seem to be trying to move to a multi-armed bandit, bayesian methodology.
Statistics is used heavily in machine learning.
Questions look answerable for someone taking 2 courses in machine learning. One being the introduction.
Machine learning systems that perform a job and make money are still overwhelmingly stats-based, and the knowledge required to understand, tune, and optimize any ml system, regardless of design, are based on statistics and always will be.
I'm not sure I wholly agree there. There's an awful lot of real work being done that makes real money based on deep learning, and deep learning comes from the CS "it works, what's the problem" side of the field as opposed to the statistics "I want formally grounded theory for everything" side.
One of the big pushes in Bayesian statistics recently has been to try to figure what the hell all these neural nets are actually doing. It's certainly not the case that the stats have been in the driving seat there.
Classical AI is applied, layered logic.
Modern ML is applied, layered statistics.
season to taste.
I have the impression that at big universities maths is always the #1 topic in CS/ML. So its no surprise their graduates ask the same riddles as their profs.
I personally know a few ML professors at my university who are not taking any CS grads as pHDs. They only want people with maths degrees.
A lot of these CS professors are themselves maths grads.
CS is discrete maths, which is not quite the same as statistics.
As long as these riddles keep the web devs and infra engineers away from machine learning, I'm all for it.
We must preserve the elite reputation associated with machine learning; not watered down by avergae engineers looking for a career change.
People have been doing statistics for decades, which is mostly all that machine learning is. Coming up with a fancy new buzz word for stuff that has been for decades, doesn't make something "elite".
You know, statistics is a branch of mathematics, and mathematical education itself can be considered "elite", if for no other reason then for the irrational fear it induces in most of people.
Most of high school graduates decides to pursue more "practical" topics like CS, leaving mathematics to those chosen few who dared to study it :)
Then why do machine learning scientists get paid multiples more than statisticians? Why do CS departments get far more funding for ML grants than the same grant branded from the statistics department?
Say what you want about ML vs statistics, the fact is: in both academia and industry, machine learning is far sexier than statistics.
Machine learning systems that perform a job and make money are still overwhelmingly stats-based, and the knowledge required to understand, tune, and optimize any ml system, regardless of design, are based on statistics and always will be.
These are good questions. A good quantitative researcher can answer most of them. They're also fairly open ended, allowing the candidate to show what he or she knows.
Interesting that it seems to be all on the stats side and less on engineering.
For anyone that wants to get the background for any of this I highly recommend ISLR (and ESL as a reference). http://www-bcf.usc.edu/~gareth/ISL/
I bet that's because statistics knowledge is much more scarce in the industry right now.
My quick and dirty answers are below. I'm thinking of moving jobs within the next year so I could use the practice. Can anyone do better?
> How do you treat colinearity?
Throw away the redundant part of the data
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
This is very problem-dependent, but it's got the potential to wreak havoc with your learning algorithms. You might get seemingly good results by e.g. always predicting positive. Think carefully about your loss function.
> How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives, profession and gender? Define a machine learning algorithm for this.
This is a messy combination of continuous, categorical, and binary data. I'd encode the data in a vector [salary, x-y co-ords, one-hot encoding of profession, and binary indicator for gender]. Something like a random forest will probably get you most of the way there. Unusual professions could mess with the algorithm, so consider grouping by industry or averaging with a model that omits the profession data.
> Is it useful to apply PCA to your data before SVM classification?
Probably, but it could be data dependent. Semi-supervised learning often improves machine learning models, and it'll lower the dimensionality of your inputs, which will make training / hyperparameter search far more efficient.
> How do you compare a neural network that has one layer, one input and output to a logistic regression model?
Not entirely sure I understand the question, but a NN is basically just nested logistic regression, depending on the activation function.
> From a long sorted list and a short 4 element sorted list, which algorithm will you use to search the long sorted list for 4 elements.
To be honest, I'd use binary search and call it a day. I realise this isn't the answer you're looking for.
> How will inspect missing data and when are they important for your analysis?
You might try to impute the missing data from similar datapoints. Maybe use something like K nearest neighbours on the non-missing components to deduce the missing values. Got to be careful doing this though, since it could massively bias your analysis: consider using a special encoding for "missing data" too.
> Estimate the probability of a disease in a particular city given that the probability of the disease on a national level is low.
This is a tough one, since there are so many ways to answer it. I think I'd start by writing down a list of factors that might allow a high prevalence of the disease locally despite a low prevalence nationally. Weather? Local wildlife? Proximity to major transport hubs? Then I might suggest a simple model like naive Bayes.
I will try.
> How do you treat colinearity?
Use an LASSO as a feature selector or regularizer.
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
User F1 score instead of accuracy as a performance metric. Try using cascade classifier.
> How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives, profession and gender? Define a machine learning algorithm for this.
More or less, the same as what you said.
> Is it useful to apply PCA to your data before SVM classification?
No, since kernel-SVM optimization is dependent on data (similarity) and not features. Also, we will be projecting the data into a basis expanded space anyways, so the dimensional reduction would be redundant.
> How do you compare a neural network that has one layer, one input and output to a logistic regression model?
An MLP-NN is the same as a logistic regressor, if all connections as 1-1 (no basis expansion), no activation function in the data-1st layer connection and Sigmoid as activation function in the 1st layer-output connection.
> From a long sorted list and a short 4 element sorted list, which algorithm will you use to search the long sorted list for 4 elements.
I am dreadfully bad at core CS questions...throw some AI questions at me. XD
> How will inspect missing data and when are they important for your analysis?
You might try to impute the missing data from similar datapoints. Maybe use something like K nearest neighbours on the non-missing components to deduce the missing values. Got to be careful doing this though, since it could massively bias your analysis: consider using a special encoding for "missing data" too.
Love your solutions. I might also try using cascade classifiers.
> Estimate the probability of a disease in a particular city given that the probability of the disease on a national level is low.
tough question. Would probably involve me asking a lot of questions about what data I have to play with. Very interview specific.
I am not a data scientist, just a researcher in cog neuro. With multicollinearity, I'd first want to know whether these variables are meant to measure the same latent construct. If so, creating a latent factor in a SEM seems the way to go. Lassos and ridge regression I've heard of, but I know very little, except thatse methods can find use when using fmri voxel timeseries...
Regarding the PCA before SVM step - PCA would help reduce dimensionality w.r.t linear correlations, kernels generally are used to obtain non-linear maps. So wouldn't it be wrong to say doing PCA is redundant? Also, while kernel SVM optimization is dominated by number of data points, computing the value of the kernel function for individual data points might depend on the dimensionality itself, would it not? Another instance of lower dimensional representation possibly helping.
Please correct me if I'm wrong, I'm still just learning this stuff.
The cost of computing similarities is definitely linear or lower, wrt. to features. But overall, the Kernel SVM will scale to the order of n^3 wrt. to the number of points. In that sense, focussing on the number of features may not make a lot of sense.
PCA isn't lossless either. We are definitely trading off some richness in the data for efficiency. When the gain in efficiency is negligible (as is the case with SVMs) the trade off doesn't look as tempting.
I am student in the area as well, so take my words with appropriate caution.
2 things I would like to add to the already very good answers in this thread:
> How do you treat colinearity?
Look at feature interactions when using decision tree based models to discover if some interactions between seemingly similar features could improve the performance of the model.
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
Don't use accuracy or other "averaging" evaluation metrics. When looking at unbalanced data you are likely interested in the minority class so I would use metrics that only look at this class for performance evaluation.
Do data augmentation for the geographic question - overlay the lat-lons with other spatial data (e.g. demographics, distance to nearest mall, etc.)
> How do you treat colinearity?
Colinearity can be handled either through variable selection, or dimensionality reduction. The right approach depends on the data. If you're trying to identify the root cause of some phenomenon, L1 regularized models will pick out variables (though they might pick the wrong one). If you just want to prevent colinear variables from taking over your model, dimensionality reduction is the right choice.
> How will you deal with unbalanced data where the ratio of negative and positive is huge?
This is one area where bayesian statistics really shines. The variance of the posterior predictive distribution for a bayesian model is inversely proportional to the amount of data you use for inference.
> How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives, profession and gender? Define a machine learning algorithm for this.
Random forests are a good answer to this in terms of predictive accuracy. The problem with random forests is that they don't give you any information about how certain your estimates are. A better technique to use would be Gaussian processes (GPs). GPs are a kernel method somewhat similar to SVMs, but they have the marked advantage that they provide a posterior distribution over their estimates, rather than just the maximum likelihood value.
> Is it useful to apply PCA to your data before SVM classification?
The right answer is that it depends on the shape of your data and the specific SVM kernel/hyperparameters you're using. If you have a square matrix and you're doing linear SVM, PCA first will be slower. On the other hand, for a highly asymmetric matrix and nonlinear SVM, PCA done by stochastic approximation of the SVD, followed by SVM will be much faster. From a prediction accuracy standpoint it isn't going to make a large difference in either case since the kernel trick effectively projects your data into a high dimensionality feature space.
> From a long sorted list and a short 4 element sorted list, which algorithm will you use to search the long sorted list for 4 elements.
To answer this properly you need to know if the list has repeated elements. If there are no repeated elements then a binary search on the first element of the 4 item list, then scanning forward from there is the correct answer. If you admit the possibility of (possible massive) repetition of elements then this is no longer the case.
> How will inspect missing data and when are they important for your analysis?
This is yet another case where bayesian analysis shines (can you tell what I specialize in?). With bayesian methods you don't have to impute a value, rather you can treat unknowns as coming from a specific distribution (which can be the exactly distribution of that variable, via empirical bayes). This prevents you from introducing bias in your model.
> Estimate the probability of a disease in a particular city given that the probability of the disease on a national level is low.
This question is kind of bad given that it doesn't put a bound on what data is available to you. In terms of model, this is just a classification problem in disguise
Panthers vs Titans Live Stream Free http://www.douglas-budget.com/calendar/event_ff8768ac-8509-1...
I respectfully disagree with everyone saying these are statistics. This is analytics. Most of them are down to earth and practical matters that I think would be lost on someone who did statistics as a math degree (but maybe this distinction is unclear because it is also observably different from straight comp sci or neural net expertise).
I recently graduated from a masters of business analytics program and would have humbly assumed that the breadth and depth of knowledge asked for in these questions would be beyond my capacity, but it was actually pretty spot on with what I had learned. Strictly speaking, I am far from an expert in statistics/math/comp sci. But these seem like reasonable questions. The best one is
"How will you decide whether a customer will buy a product today or not given the income of the customer, location where the customer lives, profession and gender? Define a machine learning algorithm for this."
This is a difficult question because for someone who has played with data science before, the answer seems to be obvious: throw the data into your favorite predictive algorithm and see the result (worst case: deep learning). What they are probably looking to see is a discussion of feature engineering, hypothetical exploratory analysis, and an effective way to present the answer (straight probabilities, weighted loss functions, outlier detection, etc.)
It appears that the current generation of "hackers" seems very keen to work for a big name company-- Amazon, Apple, Microsoft, Google, Facebook. As someone who has worked for many startups and several of the above named companies, please let me give you a bit of advice:
- A big company on your resume has no outside value to your prospects.
Seriously. There's not "hey well he worked for amazon so we know he's good" free pass in the future for having Amazon on your resume.
There might be a "well, she wasted 3 years at amazon and we all know that's a hellhole so we can get away with abusing her" signal though. (Amazon is a hellhole, I spent 2 years there, in retrospect I would have been better off walking the day it became clear they were using cult methodology.[1])
It took me many years to find my calling. Once I did then my career direction was set. I'll only be working for companies in this industry going forward, because it really is my calling.
Take a job that maximizes your creativity and ability to contribute.
Don't chase big money or big names. The two almost never go together, anyway.
[1] Magical Phrases are a key indicator of cult methodology. Also known as "thought terminating cliches" they are multi-purpose sayings that are used to invoke company ideology and shut down dissent. At amazon there are things like "It's day one!" and the like. Every time I've seen these cult signals at a company it turned out to be a toxic work environment.
A good question to ask in interviews is if they company has any slogans all the employees know.
My experience is completely different from yours in regards to every single point you make. Working at a big company (amazon) has increased my salary significantly, made me better technically, and noticeably made it easier for me to find a job elsewhere afterwards (this could be because my technical skill were better afterwards, and I had more experience). Additionally, I actually loved my team at Amazon. My manager was incredibly empathetic and caring, far more than any managers I have had at smaller companies before. I just wanted to provide a different perspective, and show that generalizing about companies, especially large ones, is a bad idea and often grossly incorrect. Your experience and the people you talked to (likely) represent a very small part of the whole company.
I had a similar experience there. There were parts of the company that seemed like they sucked, but my experience was only positive and put me in a position to learn a ton from strong engineers while building my skills up and making a good amount of money. When I started, I made less than some of my friends who chose startups, but two years out of college I was making more than almost anybody I knew.
I might not have worked on my passion in specific there, but I had a great time, learned a ton, and IMO set myself up for a great trajectory in my career.
Like most other large companies, YMMV depending on where you're at.
This is terrible advice.
First, because I think you're factually mistaken - for a lot of people, a big name on your resume is a huge, positive flag.
Second, because there are other effects of working at a large company which you're not considering. Most importantly - they're large! That means you're going to interact with a huge amount of people. In 3 years at a BigCo, you'll likely have contacts not only in various departments, but probably in several other companies as well, both because of work with outside companies, and because of interacting with employees who left.
Not to mention, you can learn a lot about the industry that you might not necessarily be exposed to while at a small startup - via interactions with those outside companies (customers, partners, etc). You can also work on seriously big things, which you won't be working on at a startup, for example.
None of this is to say that working at a BigCo is the right move. I actually tend to think the opposite, depending on various factors. But it's definitely not an all-downsides proposition.
> - A big company on your resume has no outside value to your prospects.
I respectfully disagree. Recruiters and hiring managers, and even individual interviewers take me far more seriously after having "Google" on my resume than they did before. It's not enough to land a good job automatically, but most interactions now at least start from the presumption that I'm competent, which makes a job search less frustrating.
I haven't worked for AppAmaGooBookSoft, but I know if we got an ex-Googler or similar at many of the places I've worked at, they'd be widely respected/feared by my peers on that virtue alone, and it'd certainly help a lot in the hiring process.
I know some well-known niche companies with highly-coveted positions that appear to make new hires almost exclusively from AppAmaGooBookSoft alumni (last time I mentioned this, a user correctly guessed one of the companies I was referring to from just such a vague description).
The reality is that most people hate hiring and have no idea whatsoever how to do it well (in the sense of bringing on competent people, not from the HR perspective of minimizing legal liability). Big names, in either education or employment history, help a lot. It's probably easy to underestimate how much they help if you haven't gone without them.
I would also venture that ex-BigNameCo is a lot less of a big deal if you work somewhere that's a hotspot for that type of company (Silicon Valley, Seattle). There it may be seen more as a sign of rejection than success, like "Why don't you still work at Google if you were Google material?". But even still, most companies are going to be swayed significantly by a big name on the resume.
> Don't chase big money or big names. The two almost never go together, anyway.
Do you have evidence on that? Virtually all data I've seen from salary sharing threads, and all compensation discussions I've had with startups suggests the contrary.
I disagree. I am tech lead at a small company, and when we look for help, permanent (rare) or contractor (usually), big software company experience is a definite plus. At least they have been vetted to some level by others first, and have probably had some decent exposure to current methodologies and practices. I really dislike looking for new help, and this experience provides a valuable data point.
Seriously. There's not "hey well he worked for amazon so we know he's good" free pass in the future for having Amazon on your resume.
The point of a CV is to get you an interview. A person who graduated from an Ivy/Russell, did a year or two at Google and a similar spell at Goldman Sachs, will be able to get an interview anywhere, for the rest of their lives.
They probably won't get an easier ride in the interview, it may even be harder as the interviewer's egos insist that they best one of the "elite", but they will never need to worry about passing a screen.
I used to be contacted by Amazon a lot around 4 years ago, to work with their data science team. At the time I was busy helping build up a startup (and a data science team), so always turned them down, I always put my people first. The startup turned out to be quite profitable, they are up and kicking nowadays, I am very proud of my time with them.
Nowadays I am looking for a new permanent position (have been freelancing for a couple of years). My wife happens to be a headhunter and knows how most HR and internal recruitment teams think and work. She tells me with Amazon in my CV many more doors would open. I don't regret turning them down, I know for sure had I accepted to work with them I would have never met my wife, who brings me more joy than any job would ever do... But I can't help to wonder...
It varies. At my company people seem to see Google on a resume as a positive indicator but I don't hear the same thing for other big name companies.
People generally make more money at big companies especially when you factor in 401k and healthcare benefits. However impact is lower at big companies for the average employee. If you want a job with a large impact and good/great money then profitable medium-sized companies are where it is at.
As for name recognition, it always helps. No employer is going to hire someone solely based on where they went to school or where they worked previously. However a well-known school or company can be the deciding factor between two candidates who are otherwise equal. My wife went to Purdue and did her pharmacy residency in Colorado. She later found out that she was competing with three other candidates for one position who were equally as good in the eyes of the residency program, and she won because she graduated from a well-known, top 10 pharmacy school.
This doesn't really match up with my anectodal experience - the number of recruiter emails I get every week has more than tripled since I started working at Spotify (after working at a number of smaller startups before).
I love it there :) go big or go home
My favorite ideology used to shutdown dissent is the No Assholes Allowed rule.
Me: "Hey how come you guys don't use source control?"
Manager: "Are you being an asshole?"
It appears that the current generation of "hackers" seems very keen to work for a big name company-- Amazon, Apple, Microsoft, Google, Facebook
Right, because BY FAR they are making the most progress, in the most promising technologies today, and they pay egregiously well.
If it isn't too identifying/personal, then what did you find your calling to be?
interesting.. I have worked both for startups and large companies. There are tradeoffs to both of course but my favorite so far is the large company I am at. I get to do things at a scale that doesn't exist at startups.
Can you elaborate on the whole "day one" thing? Sounds interesting.
Let me ask you this, have you worked at any of those large companies?
They clearly said they worked at Amazon.
I think its already been iterated on by other commenters, but almost all of this advice is wrong. Name brands play a huge roll in careers, even in programming ones. If you have worked a long career and havent worked at a big name company, thats a huge negative signal.