Hacking Diversity in Software Engineering and Technology
As a straight, white male, I've admittedly never given much thought about diversity in technology when I initially started programming. Mostly, I was just trying to keep my head above the water as a self-taught junior software engineer. I did some small time web development and systems administration, but my first true industry job was when I consulted for a technology division of AIG. It was a two hour commute, and I was the only white American programmer. Out of about 50 developers on the team for the project, there was me, two Russians, one Chinese woman, three Pakistani men, and the rest were Indians. This was in 2001--not long after the dotcom bubble had burst. The team was certainly heavy on the testosterone, but we weren't devoid of female programmers, and all the technical leads were women. This was during a time when many with H-1B Visas were fighting to say in the U.S., and consulting companies were undercutting each other to keep their people in the pipeline. This meant cheap foreign labor for companies like AIG, and for me, it meant that my initial exposure to the software engineering industry was one of diversity in the enterprise.
This likely colored my perception. Don't get me wrong, there was certainly prejudice, and plenty of side-eyes. There were a lot of American developers angry that so many jobs were going to Indians. I don't doubt the racism inherent in many industries, but I never saw the implicit bias against diversity--whether because my early experiences didn't match the norm, or because my white privilege masked it from me.
Clearly, we have a problem.
Women are not only underrepresented in technology, but underpaid as well. Many are made to feel uncomfortable due to a boys club mentality in many programming circles, and start-up culture--especially in Silicon Valley--is rife with tales of the brogrammers. Meanwhile, you can't go 10 tweets on a programmer search on Twitter without seeing a tweetstorm arise from some dude-bro attempting to mansplain Java to the original female poster, who probably has 15 years in the industry.
What about African-Americans? Nearly non-existent in a lot of programming circles. In fact, just about any non-white ethnicity prevalent in America seems to just give up on taking up the keyboard as a craft.
When I started working at the University of Virginia, School of Medicine, I was quickly placed in a position where I had input on hiring decisions. I watched as applicant pool after applicant pool consisted of the same statistics. Out of about 10 applications, there were maybe 3 middle-aged white guys who have spent 15-20 years doing the same thing over and over again, and the rest of the queue consisted of Indian or Nepali developers fresh out of college, trying to find a way to maintain their Visa status. In nearly three years at the School of Medicine, I've only seen one female programmer in any of the pools, zero African-Americans, zero Latino/Latina individuals, and nobody openly a part of the LGBTQ community.
This isn't for lack of trying. In the last several hiring rounds, I made a concerted effort to work with organizations made up of people from underrepresented groups. I utilized Twitter, LinkedIn, and other avenues of social software to spread the word. I even contacted individuals directly who I knew were solid programmers. Despite all of these efforts, our applicant pools remained the same.
Recent events surrounding the Robert E. Lee statue in Charlottesville didn't help. I even had a Jewish woman reject an interview because of "everything going on over there."
This isn't representative of Charlottesville either… and not even the industry as a whole. When we took an entry level web developer position and changed the title to "tester," the applicant pool filled up quickly with women. There are societal considerations that make things more difficult.
Psychologically, women seem more likely to forgo applying for a position unless they meet all of the requirements. Men, on the other hand, will routinely apply for jobs they aren't qualified for. Women are also more likely to accept the first offer made to them, while men are more likely to negotiate. African-Americans, meanwhile, have been talked down to by society so much that their anxiety levels increase in situations where cultural bias has decided that they are inferior--including tests, interviews, etc.--despite the problem truly being implicit bias.
This is a big problem because it leaves the design of software in the hands of one group, capable of altering cultural norms even further. If straight, white men can't spot biased data sets that are being used in supervised learning, our neural networking models are going to incorporate that bias into their probabilities. In fact, data science and machine learning today faces the same issues that psychology did a few decades back. Psychology had a problem where most research, test construction, and case studies where done by European (mostly Germany), American, or Canadian researchers. That's really white. This led to culturally biased psychological and educational tests, as well as the generalization of research that wasn't representative of non-white populations. Psychology has attempted to solve its problem through injecting multicultural sensitivity in much of its college course work (although there is still more to do). Computer science would do well to follow suit.
Software in the hands of one specific group leads to software that is difficult--and sometimes useless--in the hands of others. It's kind of like when Apple's Health app launched without support for tracking menstruation. Difficult and useless is one thing, but this can easily crossover into being dangerous when that software is used to determine things like jail sentencing.
Diversity helps improve the utility of software, and by making software more useful for even more people, we help make the world a better place.
What's the answer? It's more than just better hiring practices. As my story emphasizes, even a strong recruitment drive can fail to yield results. We need to make sure that outreach occurs in the community on a regular basis, and that outreach needs to be at the middle school, high school, college, and professional level. It's just as important to inspire that 40 year old looking to switch careers as it is to inspire that sophomore in high school.
In addition, we (as straight white men) need to be careful not to assume that those in underrepresented groups need our help or mentorship. Cornel West posited in his essay A Genealogy of Modern Racism that the assumption that African-Americans need to be helped and elevated to the same level as white Americans is an assumption that African-Americans are somehow not on the same level as white Americans--somehow beneath white Americans. This is not the case. African-Americans do not need to be elevated; they simply need opportunity. In the same regard, programmers from underrepresented groups are not necessarily in need of mentors; they are in need of opportunities. They are in need of the exposure of unintentional bias, and the elimination of explicit roadblocks.
In the Fall, the University of Virginia's Provost office put out a call for proposals to fund initiatives meant to help overcome intentional and unintentional bias. I submitted a proposal to host a hackathon exclusively for underrepresented groups. On April 14th of this year, we'll be holding that hackathon at the Jefferson School's African-American Heritage Center in association with the Tom Tom Festival. It'll be an all-day event with free food, door prizes, swag for participants, and Visa gift cards for members of winning teams. Beyond the prizes, this is an opportunity for members of underrepresented groups, who feel that they are marginalized in software engineering and technology, to network with each other, and be inspired by each other's stories. My hope is that everyone has fun, makes friends, and comes out of the event a little more confident in their ability, and knowing that the University of Virginia--and the technology companies of Charlottesville--want them to be a part of their respective teams.