Ask HN: How do I get out of data science?

[StClaire]

I spent a few years trying to land a job as a data scientist, and now I want to do something different in the tech field. I haven't had luck finding a new position :(

I've looked at data engineering, systems engineering, signal processing, project management, and architecture roles. I've done some of each of those in my current position and I've enjoyed them much more than the machine learning I've done. But I also don't feel like I know enough of any of them where I could make a lateral move and have any degree of success.

When I've applied to jobs, I've sometimes gotten an interview for the role I applied for and got one job offer that turned out to involve a substantial pay cut. When I hear back, I feel like most of the time the person who calls me wants me to work on their data science team and then gets kind of confused (and sometimes offended) when I tell them I don't want to work as a data scientist anymore.

How do I make this transition? Is it just a matter of time or do I need to approach things from a different angle?

[lacker]

Think of it less as "I need to get out of data science" and more "I want to get into X". Because the answer is going to be pretty different about how to get into, say, systems engineering or project management.

If you are good enough at software engineering to do a generic software engineering job, then I suggest applying to one of those in the standard way. It is much easier to get into all of those areas once you have some software engineering experience.

If you are not good enough to get a different job right now that would satisfy you, then I suggest talking to your current employer about moving your responsibilities in a different direction.

[st1x7]

First you really need to decide what you want to do next. The roles that you've looked at vary a lot and some are easier to transition into.

- Data engineering? Could be easy to switch to but to some people it would look like taking a step back. (I don't think so personally)

- Signal processing? Depending on the type of data science work you've been doing, it could be almost indistinguishable from your experince so far and very easy to transition into.

- Systems engineering? Probably the most difficult one out of your list and it will take a lot of additional effort and studying.

- Project management? It's a completely different career direction but you can probably gain some experience with it while you're at your current job and slowly make the transition if that's really what you want to do.

- Architecture? It helps to have a lot of experience as an individual contributor.

Pick something and make a plan for getting there, instead of applying to a random mix of jobs just so you can get out as quickly as possible.

It's also good to be honest with yourself about the reasons for wanting to leave data science. Depending on those reasons it's possible that you just need to change your employer instead of your field of work, especially since data science can mean so many different things in different companies. If this is your first data science job, don't assume that the field is limited to whatever your current employer happens to label as data science. I can think of more than a handful of jobs that are very distinct from each other, all of which can have the title data scientist.

[dasboth]

You say "I've enjoyed them much more than the machine learning I've done". I actually find the machine learning the LEAST interesting thing about a data science career in the sense that it's just another tool in your DS toolbox to solve open-ended problems that no one has properly tackled before. It sits in the same box as bar charts, web scraping, asking better questions etc. and you can do valuable work with a variety of approaches, not all of them machine learning. Not sure exactly what part of data science you feel you don't enjoy, and it sounds like you DO enjoy a variety of technical disciplines, but I was just curious whether you wholly equate data science with machine learning.

[blt]

"Data science" can imply many different roles, but one thing they have in common is the absence of any low-level component. They usually work with high-level garbage-collected languages on powerful machines and reliable LANs.

I am not a hiring manager, but if I saw a resume with only "data science" roles, I would assume (unless indicated otherwise) that they don't know C, can't read ASM, can't understand a wireshark trace, etc.

If you are not getting to the interview stage for roles that include this kind of work, it may be because your resume fails to disprove this hypothesis.

[mssundaram]

Can you describe in more detail exactly what you're looking for instead?

[paulbishop]

stop showing up