Hacker News Reader

Comment by randomacct3847 | Hacker News Reader

I’d wager 90% of working developers with a CS degree also could not pass a test in some random mathematical topic they studied X years ago in college either.

Also, while there are those that go from bartender to developer via bootcamps, a lot of people who do these camps have technical undergrad degees (other engineering, math, physics, etc.) and probably took as many math classes for their major as a typical CS grad.

And let’s be real...most developers, maybe especially at FAANGs, are frankly overqualified for their day to day job and aren’t doing much more than things they learned in a CS 102/103 class.

View on HN

I think I understand what you're trying to say, but I don't think what you've actually said is correct. I'm a >senior engineer at a FAANG and have been in the software game for more than ten years now. I definitely do not feel overqualified for my job, and I don't know anyone else who does. Or, at least, I don't know anyone who actually is by their demonstrated skill at their work. I guess I can't know their internal feelings unless they tell me.

What I can agree to is that the job has significant challenges that a CS education does little to prepare you for. The interpersonal challenges. The challenges of designing large systems, forward and backward compatibility, the interaction of business and technical priorities, deployment, production safety, etc. And it makes little use of a number of skills that were very important to CS performance. I can tell you I've used zero graphics programming knowledge, little of my compilers course, very little of what I learned in my networks course (despite that I'm regularly called upon to write programs that communicate over the network). I've written exactly one proof since I started my professional career. In that sense, my CS program over-qualified me for the work. But I think it's more accurate to say that I'm differently-qualified, education-wise, from the field I'm actually working in.

Replied to the other comment, but I hear too many stories from angsty mid-career FAANG developers who do little more than write a bunch of Protocol Buffers all day, every day.

That isn't the type of work I do.

I wasn’t suggesting that you did, just saying anecdotally hear a lot of gripes from mid level FAANG developers whose day to day is that, hence the point that many, but not all, developers are probably overqualified for their day to day role.

I don't know about overqualified, it's more that the job is hard in a different way. I'm not a CS grad, but it's not as if I'm actually using my numerical analysis or computational fluid dynamics coursework either.

However, I do jump between discussing if a trie is the right way to implement something, discussing whether the "simple" model should be a SVM or boosted tree, trying to figure out if we've covered the major abuse angles in our product, coordinating with all the other teams we're collaborating with, analyzing the experiment results, etc, etc.

Having at least some actual CS certainly helps fill in the gaps, even if it's not what I'm doing most of the time.

Do CS grads study svm?

Yup! Senior computer engineering student, we covered SVMs in our machine learning 2 class!

Many CS programs have machine learning courses, but at least when I went to college they were not required.

> I’d wager 90% of working developers with a CS degree also could not pass a test in some random mathematical topic they studied X years ago in college either.

I wouldn't, either. But I've discovered that if I go about relearning it, it comes back very quickly. Having learned it before, I also recognize when I need to go relearn it, and what I need to know.

It's a lot better position to be in than having no idea how to go about figuring out how to solve the problem.

So true, it's the unknown unknowns that tripped me up before going back to get a Masters CS degree.

At FAANG most of the work is boring copying from one datastructure from another. The remaining 5% needs understanding on working on big data, which means statistics (as long as you're working on backend...I'm sure frontend needs less math).

I haven't seen differential equations, but sometimes calculus, probability theory and a basic understanding of machine learning and deep learning is needed to understand what other teams that you're working with are doing, or to analyze the statistical significance of an experiment that you ran.

At FAANG I often wished I knew probability theory and statistics better (instead of other things that we learned at my University).

> maybe especially at FAANGs

It's hard to say how it is in the large, but at Microsoft I do apply CS principles fairly regularly, use statistical tools, and think fairly regularly about different data structures and algorithms. A lot of this is due to working in performance-constrained environments. If that were not the case, then I would say the academic side of things would be less prominent.

When I was at Microsoft I didn't apply CS principles beyond occasionally thinking "This function is going to be O(n^2), maybe I should change it. Oh, we'll never have more than 100 or so items, so it's probably not worth fixing that."

I should say many, but not all.

I hear too many stories from angsty mid-career FAANG developers who do little more than write a bunch of Protocol Buffers all day, every day.

Reply on news.ycombinator.com