Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: [Data Science] What would make you hire a guy who self studied?
6 points by WaffleSmack on Nov 23, 2016 | hide | past | favorite | 10 comments
What would you need to see on the resume of someone who is self studying to become a Data Scientist if he didn't have relevant degrees (like no Math or Com Sci or anything of that sort)? A lot of people have recommended a plethora of resources online, but which ones when you see on a resume (assuming you even put those in a resume) would interest you to bring him in for an interview?

I roll in an unusual way, so take this with a grain of salt:

I'm very dubious of resumes. Typical resume is an alphabet-soup of "hot" technologies alongside a 4.0 GPA they got by remembering, regurgitating, then forgetting.

Two things that seal the deal for me are: A) well-executed personal projects that they can B) discuss their methodology fluently, casually, and competently so that I know it is truly their own work. If they can do A and B, I don't even care if they graduated high school.

That, and the definition of "data scientist" is somewhat up-in-the-air. A lot of jobs listed as "data scientist" are really more like map-reduce DBAs than anything remotely resembling science.

While not data science specific, I do often find myself on the hiring/search committee for our tech related jobs.

And honestly? Aside from the obvious "experience" (at least a year of contiguous experience doing something similar to what we want), there really isn't anything that would make me give a checkmark to a fresh faced person without a remotely relevant degree. Because online courses are useful but don't really show me that they know how to work on a project as a team. Same with most github projects where there is one contributor.

Universities may not be great, but it is pretty hard to graduate without ever having worked on a group project. Moreso for advanced degrees. And that is what I want.

Obviously there are exceptions on both sides. But when we have a hundred CVs to sift through, we are going to narrow it down. Because none of us want to do 100 interviews. It isn't fair to us and it sure as hell isn't fair to anyone past number 10 where we genuinely won't care anymore.

Thanks! It was good insight

No problem. And good luck. We are a fairly established firm so our approach is more "We need someone competent who can work well with others". It sounds like start-ups are a lot more willing to take a risk and think outside the proverbial box.

Which I don't at all get as one would think we could absorb a failure or a delay more, but different cultures.

Here is the path I'm following:

I'm doing the Data Science Specialization from Johns Hopkins University / Coursera, with verified certificates that I hope will help me create a portfolio to showcase as I look at this type of work.

I feel like having a portfolio to point to, as well as code on a site like GitHub, should be a good basis for a conversation with a potential employer.


Demonstrate that you can turn data into Money.

The how is intentionally left open. As a scientist,you need to be capable of owning the entire process. You need the hustle to acquire your dataset and the mathematical ability to derive an advantage from it.

Can't be learned at school. You have it or you don't.

It's OK if you aren't able to "realize" the making Money part due to real world constraints. (For example a demonstrable investment opportunity that you can't personally invest in because you don't have funds to play with)

Thanks! Good reply

Info on their blog (or first phone call) that proved the person knew fundamentals, and some insights myself or team found interesting and fresh.

Your best bet would be to get experience in a software engineering job that works closely with a data science team and weisel your way in

Maybe some projects and/or relevant experience in the field

Applications are open for YC Summer 2021

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact