Civil engineering PhD student wanting to transition to a data science role

Question

I am graduating with my PhD in civil engineering in the next year. Considered staying in academia once upon a time, but wanting out of it for various reasons. Mostly unhappy with the overwork culture, relatively low pay, and little time off.

I know how to code in Python/matlab and have developed a lot of my own codes to analyze large data sets. I know I’m missing machine learning experience, but have taught myself some (plus SQL).

It would be great to get some of your thoughts/advice on a few questions
-How will my lack of machine learning experience affect me getting hired?
-How less desirable is a PhD in CE over one in mechanical, electrical, chemical engineering for a DS role?
-If applicable how has your experience been as a woman in DS? Is there still a lot of workplace discrimination or issues getting hired bc of this (I’m based in the US)
-Should I just be studying a lot of Leetcode questions to make sure I can pass interviews?
-How is the job market for DS roles currently?

Other advice welcome!

Answer 1

Many people don’t have a degree in STEM, let alone a PhD. Your academic credentials are not 100% perfect, but 97,5% perfect. You’re academic background won’t be a problem at all.

Machine learning is obviously important, but it’s not even 10% of job tasks a typical DS has to work at. What is more of a problem is a lack of industry experience in any relevant field (like finance, generic IT, etc).

Regarding gender, most probably it doesn’t matter much. I spent some time hiring and interviewing people, gender was never a factor. Though, your resume probably has a higher chances to be chosen for the first interview. But after that it’ll be your knowledge and your performance on the interview rather than your gender.

Leetcode might be useful for some programming heavy roles, but to be honest 95% of applications don’t require that much in terms of coding. So, I’d spend time grinding leetcode only if you plan to work with computer vision and/or in MLE roles. For “classical” DS it’s unnecessary.

Learning solid SQL, BI development, ETL skills might prove more useful. Many nominally data science positions are very light on coding and modeling, and very focused on data visualization, BI and data analytics.

Before getting to DS you need to understand that with 95% probability you won’t be creating ArRiFciAl iTelLiGenCe, you will be a glorified Business Intelligence Analyst with robust stats and coding skills. If it’s fine with you, learn Tableau, Power BI, ETL, Excel VBA skills, as I said above - they’ll land you on a job.

If “AI” development is what you’re looking for, those are MLEs that are doing it nowadays, and to become one you’ll probably need time as a SWE and/or DS/DA.

As for job security during recession - it’s kinda meh, if DS department is a breadwinner, they’ll probably be ok, but many data scientists are rather tertiary support positions just helping business, barely creating tangible value. Those people are fucked, haha.

Answer 2

I was in a similar position to you after grad school, I got a master’s in a physical science and had 2 years of experience programming in Python and Matlab, working with huge datasets, and building custom algorithms. I had done a Coursera class on machine learning and liked it, and I didn’t want to stay in academia so I thought data science would be a good fit. I applied to jobs for months with little results because I didn’t have much machine learning experience in my resume. So I decided to do a 3 month DS bootcamp - it was expensive, but it did pay off immediately. The best part was they required us to do a bunch of ML projects including a personal capstone project that we could put on our GitHub and link in our resumes to showcase our experience. I had 3 offers for Data Scientist positions within 2 months of graduating. It was a huge benefit to have a masters, and I imagine PhD is even better. Members of my bootcamp cohort that only had bachelor’s degrees (even in a science field) had a much harder time getting responses.

In terms of being a woman in DS, I have had no issues with getting hired or discrimination. I would study some Python Leetcode and SQL because it definitely comes up (especially in interviews with bigger tech companies), but make sure you have at least 2 ML projects you can talk about as well. Many of my technical interviews were just the interviewers wanting me to walk them through my capstone project and asking me why I made the decisions I did, what the challenges were, etc.

Answer 3

I have a M.S. in civil engineering (plus 4 yoe) and recently transitioned into the data science field. Given your background you shouldn’t have a difficult time picking up the necessary skills and finding a job in the industry. The biggest thing you should work on is building out a solid portfolio to showcase your skills.

You might want to consider expanding your search to analyst roles as well. Once you have that initial data related job, it’s so much easier to jump to a place where you can get the data scientist title.

Answer 4

Just wanted to say I'm in the exact same boat as you! Graduating with a PhD in environmental engineering this May and already so disillusioned with academia. I can't imagine grinding it out to make $60k for the next 3+ years as a postdoc just to most likely not get a tenure track job.  good luck with your career transition! I'm doing a coursera ML course right now with the goal of putting together a capstone project to add to my portfolio.