Boot camp for snowflakes

May 11, 2015 12:51 PM

Transitioning out of academia, considering trying to rebrand myself as a "data scientist," wondering if a boot camp or certificate program would help. Snowstorm inside, but there's basically two issues. One: Is a program like that a good way to deal with having background and skills that don't show up on my resume? Two: Is a program like that worthwhile for someone who… let's just say "won't fit in culturally or demographically at tech industry networking events"?

I'm trying to keep this as brief as I can, but I'm happy to give more details if you think they'd be relevant.

1. I spent a lot of grad school (Ph.D. in linguistics) doing what is now called "big data"/"data science." Worked with datasets containing billions of words, did lots of stuff correlating language use with demographics, played with various NLP and machine learning tools, tinkered with data visualization. I'm no software engineer, but I can get by in R and Python and have some basic Haskell ability. Etc. etc. etc.
2. But! My resumé, portfolio, etc, make me look like someone with a humanities or lite-and-fluffy social science degree. (During the time I was doing the "big data" work, my personal life was a fucking mess. By the time I got my shit together, I'd burned all my bridges in that lab. I switched advisors, and as a result finished my degree doing unrelated stuff — field work on a minority language. I have no publications or even finished projects that draw on any sort of NLP or data-crunching techniques.)
3. I'm interested in a career in tech or data science — not just because that's where the money is, but because playing with data is a lot of fun.
4. I'm a visibly trans woman in my mid-30s. I don't drink, don't play computer games, and don't care about media fandom. I don't live in the Bay Area, and don't want to. (Right now I'm in Boston.) Startup people who meet me instantly write me off as Not One Of Us. One big advantage of bootcamps is supposed to be access to a professional network and to recruiters, but I worry that that will be wasted on me.
5. I have applied for data scientist/data manager/etc jobs, and occasionally gotten through a round or two of phone screening, but have no offers. I suspect my lack of credentials is part of the problem. (I worry that my tits are also part of the problem, but I'm not giving those back.)

Under those circumstances, would a data science bootcamp or a professional certificate in data science be worth pursuing? Are there specific programs I should look at, or specific ones I should avoid?

posted by nebulawindphone to Work & Money (14 answers total)

One possibility to consider: build a portfolio of data science projects that you can refer to in your resume and interviews. Host the code and analysis at Github and the raw data on Dropbox.

I can't help with any other part of your question.
posted by dfriedman at 1:07 PM on May 11, 2015 [1 favorite]

Best answer: If you did it and it's not showing up on your resume, your resume is the issue. Industry resumes are not CVs, and the vast majority of us in industry don't measure ourselves by publications. We measure ourselves by projects we worked on and results we achieved. You need to have accomplishments by industry standards in order to get a data crunching role. Think:

- Identified N new correlations in billion record data set to support lab research and publications.

I'd think you'd be an ideal candidate if you positioned yourself correctly -- having experience with the work, having team experience in the labs, and having excellent communication skills from the social sciences.

I would highly recommend the Manager Tools interviewing series for upping your interview game and better understanding the corporate hiring process.

I would also have a more professional explanation of why you left your lab (if it comes up!) than "my personal life was a fucking mess." It's not a death sentence to have failed -- hell, I've been fired. But you need to be able to confidently say "Hey, I made a mistake, I moved on, learned from it, and have been happy to put that behind me."

Some other thoughts: take a look at the job descs for the positions you're getting callbacks on, and those you're not. Is certification the difference, or something you have but aren't communicating on your resume? I'm willing to bet it's probably the latter if you're already making it to round two.

And regarding being a trans woman, I'm not going to presume to know what that's like, but I can say this: if a company won't hire you because you're a trans woman, working there would suck anyway. The company you want to work for is going to care if you produce results. So make your resume about results, don't sweat the no's and the no callbacks, and find that place that doesn't suck to get your start.

You're closer than you think!
posted by bfranklin at 1:08 PM on May 11, 2015 [12 favorites]

Best answer: Don't get the certificate. It's throwing money at a problem that money might not really solve. Instead, I'd focus on getting some publications based on your big data work out. You don't need to be in a lab to publish. Good luck.
posted by sockermom at 1:28 PM on May 11, 2015

At least in other parts of the tech world, there is a lot of skepticism directed at bootcamps and certificate programs. While it is possible that data science is an exception, I rather suspect that it is not, and your time would be better spent polishing your resume to better reflect your actual experience.

If you spent time doing the work and learning the skills, and it's relevant to what you want to do, it should be on your resume. The fact that you didn't graduate from that program isn't relevant. You may need to figure out how to answer questions about your transition from that program/lab to the other one from which you actually did graduate, but that's a different issue. That stuff needs to be on your resume if you want to do that as a job. Otherwise can you really blame someone for not considering you for the position? They have no way of knowing you've done all this stuff (R, Python, Haskell, etc.).

So put that stuff on, don't short yourself on your resume of experience you actually have, and see if that gets you more callbacks than you're currently getting. Putting together a portfolio of work—maybe not academic papers in the formal sense, but projects, code samples, heck maybe even technical blog posts—might help both to keep you interested / engaged during the job search and also give recruiters something to look at.

And there are lots of tech jobs on the East Coast; SF is overrated IMO unless you want to live a very specific lifestyle.
posted by Kadin2048 at 1:48 PM on May 11, 2015

Response by poster:

Industry resumes are not CVs, and the vast majority of us in industry don't measure ourselves by publications. We measure ourselves by projects we worked on and results we achieved. You need to have accomplishments by industry standards in order to get a data crunching role. Think:

- Identified N new correlations in billion record data set to support lab research and publications.

Yes, I'm well familiar with this advice, and I already do this on my resume wherever possible. So my fieldwork experience, for instance, is described not in terms of publications — because who cares — but in terms of grant money won and managed, and research assistants supervised, and travel logistics planned, and so on.

It may just be that I'm being unimaginative, but I can't see how to reframe the data work I've done in terms of industry-style accomplishments. ("Identified N new correlations"… doesn't even seem like a meaningful metric in this context? I honestly don't even know what that phrase would mean here.)

One possibility to consider: build a portfolio of data science projects that you can refer to in your resume and interviews. Host the code and analysis at Github and the raw data on Dropbox.

Instead, I'd focus on getting some publications based on your big data work out. You don't need to be in a lab to publish.

I no longer have access to the data that work was based on, or permission to use it. In some cases, putting it on the open internet would be illegal.
posted by nebulawindphone at 2:10 PM on May 11, 2015

Best answer: It may just be that I'm being unimaginative, but I can't see how to reframe the data work I've done in terms of industry-style accomplishments. ("Identified N new correlations"… doesn't even seem like a meaningful metric in this context? I honestly don't even know what that phrase would mean here.)

Unfortunately, I can't provide more detailed recommendations because it's outside my area of expertise.

but in terms of grant money won and managed, and research assistants supervised, and travel logistics planned, and so on

Grant money won and managed is great for an accomplishment, but I'd caution that the other two are responsibilities rather than accomplishments. My go to rule for these things is "what have I done that would be better than just another warm body in the role."

If you'd be interested, I'd be willing to look at your resume in conjunction with a specific job desc and give you an assessment of where I think the issues are. My email address is in my profile.
posted by bfranklin at 2:21 PM on May 11, 2015

As far as a coding portfolio, would it be possible to whip up a script relatively quick to make some "test data" that scripts/codes you've written could analyze?
posted by Zalzidrax at 2:35 PM on May 11, 2015

Check your memail. Tl;Dr -- everywhere is desperate for numerate people who aren't afraid of coding. Camps should be paying you.
posted by gregglind at 5:09 PM on May 11, 2015

Get new data and do things with that and submit papers to peer-reviewed journals. Use Twitter to find potential collaborators, if you have an academic Twitter presence.
posted by sockermom at 5:22 PM on May 11, 2015

Best answer: What kinds of networking events have you gone to? Would you be interested in something like Code for Boston? It would be an opportunity to meet people and build your portfolio. I don't think I'd recommend a bootcamp; I'm a successful graduate of a web development bootcamp and even I think data science bootcamps are scammy.

Also look into some of the women-focused networking groups. Pyladies is a focused on Python and specifically describes itself as being for all who identify as women or genderqueer. I don't go to that group but I do attend the Boston Ruby Women group sometimes and I don't know if it's as explicitly trans-friendly but there are definitely some queer ladies among the regulars.
posted by mskyle at 6:00 PM on May 11, 2015 [1 favorite]

Best answer: I'm a female data scientist working in the Boston area in a tech-adjacent* field. I'm not in charge of hiring so I don't see all the resumes we get, but based on the candidates that my boss decides to bring in after phone interviews, I'd have to think that your background would make you reasonably competitive, at least in our pool. (Our last hire was a physics PhD with some experience in machine learning, R, Python, and so on, and little to no non-academic work experience.) To echo others, I think you will be able to make this work.

Frankly, at this point I'm not convinced that boot camps/certificates/what have you in data science are hugely useful. The field is in a lot of ways so young that they haven't really had a chance to prove their worth. I'd focus instead on making your quantitative work visible, whether that's through descriptions of your grad school work or work on open data sets.

I'd be happy to talk via MeMail about any of this in more detail. (And I for a while seriously considered graduate school in linguistics, though not NLP, if you think background on that half of things will be useful.)

*When I say 'tech-adjacent', I mean that my department and the ones I work most closely with operate with some of tech attitudes (I swear a lot and wear jeans to work), but the company as a whole isn't tech/startup. This means we have a culture where my co-workers (both men and women) say things like 'I'm going home to my kid, this problem will be here in the morning.' I also don't find hardcore startup culture particularly appealing, and this is a happy middle ground for me - there do exist tech/data jobs that aren't all startup all the time.
posted by anne_severson at 8:54 PM on May 11, 2015

Best answer: If you enjoy programming, and decide that you would like help finding a job in big data, or even just time to brush up on your skills, I'd strongly recommend looking at Recurse Center. The key thing is that it's not a bootcamp - they like to describe themselves as a "writer's retreat for programmers." You show up every day to be surrounded by friendly, interesting people and just a little bit of structure to prod you to work on your own project. In your case, I picture you using those 12 weeks to create a portfolio showing how you can analyze large datasets. Past recursers have made some pretty cool visualizations of public transit data, political data, etc. The only restriction on what you work on is that it has to be open source; putting it in your public github account is sufficient, and exactly what you want to do anyways for a portfolio.

Relevant points:
* They are explicitly trans* friendly. If you're interested, I can probably find a trans* alum who is willing to chat with you about their experience there. Memail me!
* Free. Absolutely free, no strings attached. Plus, as an underrepresented minority in tech, they'll give you up to a $7k grant if you're admitted and can't attend without financial assistance.
* They make all their money from recruiting fees, and would like you to take a job through them at the end of your batch. However, when I attended, there was no pressure to do so. The founders are absolutely amazing, and I've watched (and benefited!) from career/interview/negotiation advice that they willingly gave even though they weren't going to get any financial benefit from it.
* They admit people based on being nice, interesting, and having a desire to improve as a programmer. I can't emphasize enough how important the first two qualities are to the admissions process. There are plenty of ex-academics that have attended, so your background would fit right in. (Heck, in addition to ex-Googlers, they've had an ex-ballerina and even at least one linguist.)

Potential caveat: Most of the companies that they have recruiting relationships with are in NYC. They're growing in the Bay Area, and I think they have some contacts in Austin and Boston. However, even if you're absolutely stuck on staying in Boston afterwards, I think it could be a good option for you.

p.s. You may have heard of them as "Hacker School" - they recently changed their name.
posted by Metasyntactic at 10:01 PM on May 11, 2015 [7 favorites]

You might want to have a look through the archives of the Geek Feminism blog, if you're not already familiar with it. They sometimes post links to tech programs and events that are geared towards non-traditional audiences.
posted by une_heure_pleine at 12:45 AM on May 12, 2015

Best answer: I work as a data scientist for a big telecom company; you'd certainly be competitive here.

As for the bootcamps: I think there's a tendency among people transitioning out of academics to be attracted to that route because, well, it's more school. What you need now is not more school. Figure out how to present the work you did in a positive light. And if you can talk intelligently about that work in interviews it's not incredibly important that people can't see it; people who have been working in the private sector have this problem all the time. We don't care so much what you did before as what you can do in the future; we'd be hiring future you, not past you.

As for not feeling like you fit in at startups: maybe you don't. But startups often don't have good data. You may actually be more attracted to working at a larger company that has more data (and maybe better work-life balance!)
posted by madcaptenor at 8:19 AM on May 12, 2015 [2 favorites]