Principal

I’m a computational scientist working at a biotech company at a level equivalent to a Principal/Staff IC at a software company. The world of scientific computing is famous for shoddy software: think one-off Python/R scripts with a single 10k line __main__() function, zero version control, and no semblance of engineering or coding rigor. While this is the unfortunate norm in most of academia and industry, the computational biology division of my company differentiates itself by eschewing this trend and acting like a real tech company. We take pride in having a very well-engineered codebase, and it’s a large factor in the company’s success in a very competitive market. The company’s customers consistently tell us that we have the best software and analytical methods in the field, which is a big reason why they use our products. The computational biology division is about 90% men. About 25% of our hires are women, but their tenure at the company is much shorter than men’s (median of 2.5 years, compared with 5.5 years for men). A VP at the company (“Velma”) was tasked with improving this attrition discrepancy, and she met 1:1 with all senior members of the division, including myself. Velma told me that the reasons women give for leaving are not the usual suspects, like bro-y culture, intellectual dismissal, outright sexism, etc. Instead, she said that the overwhelming reason women are dissatisfied is our focus on “engineering minutiae” (her exact words). She gave an example of an interaction I had with “Susan” on our team. Susan wrote a tool that used O(n^(2)) memory, which worked fine on test data but blew up on real data. Rather than implement a simple algorithmic fix that would let it run in O(n) memory, Susan’s solution was to just provision a VM with a ludicrous amount of RAM (>1 TB). I was responsible for reviewing her code, and she pushed back when I told her this would be unacceptable for production use. (Her pushback was along the lines of “the biggest AWS VM has 32 TB of RAM, so until we hit that I don’t see any problem.”) Furthermore, according to Velma, Susan was actually very upset that I asked her to implement the O(n) fix, feeling that I was “trying to run circles around her by showing off my knowledge of obscure CS trivia.” That said, Susan did not directly voice this displeasure to me, and with some guidance, ended up implementing the fix. Her tool now runs great in production. My 1:1 with Velma was eye-opening. Thinking back, there is a definite pattern of women on the team writing code that is generally scientifically sound but poor from an engineering/CS standpoint. I did not realize that women specifically were consistently being put off when asked to address these problems. (The opposite problem crops up with some men on the team, whose code is overoptimized and overengineered to the point of unmaintainability. From what I can tell, they are not upset when asked to simplify things — the worst reaction I heard was something along the lines of “that was a bloody clever piece of code and it’s a pity people aren’t willing to take the time to understand it.”) Velma agreed wholeheartedly that we would not change our rigorous engineering standards, and that there is no quick-fix to this problem. She just asked that I be aware of it, and reflect over the coming months over potential ways we can address it. Given the fairly nuanced and levelheaded takes I’ve seen here on gender issues in tech, I thought I’d ask this sub for any advice or experience. Thanks so much! Edit: Thanks for all the great replies! Lots of things to think about. One common thread I want to address: I've seen several comments saying that this is jumping to conclusions based on a one-off anecdote. I only listed the Susan story as an example; Velma gave several other such examples, so she's not basing her conclusions on a one-off. Velma is being extremely rigorous about identifying this as a systemic problem; she went through transcripts of all of the division's exit interviews over the last few years, and interviewed multiple current team members.