In the technology industry, we talk a lot about the lack of women in technical or leadership roles. In cybersecurity, the issue is even more precarious and dire. As you might guess, things get worse as you move into more niche positions. In data science, for instance, it can be even tougher for a woman to break into the business, no matter how talented and qualified she is.
While it’s thrilling and inspiring to see all the work going into getting more young girls into science, technology, engineering and math (STEM) programs, we also have to also consider the unconscious (and, sadly, probably conscious) bias of leaders and recruiters seeking new talent to fill technical roles.
In my experience, it’s rather difficult to be taken seriously as a woman doing research topics involving mathematics and statistics. There is a general unconscious bias that many of us have, myself included, that tends to take masculine people at their word when it comes to analytical results, where we may more likely question the same results if presented by a woman.
The assumption goes that men know what they’re talking about, whereas women might not fully grasp the numbers, algorithms, and complexity they’re dealing with. It’s usually unconscious and entirely unintended, but it’s a real, measurable effect and it IS sexism.
In my own career as a technical female, I have seen many examples of this, but one in particular experience in my past makes for an interesting case study.
As part of a larger machine-learning study at a previous employer, I needed to solve and then hand-code a rather difficult piece of calculus for my machine-learning tools to consume (a messy gradient problem, for those of you who are interested.) I did the math, and I wrote the code, but the algorithm wouldn’t work.
My team leader said I must have coded something wrong. I thought he was probably right. I’d checked exhaustively for bugs and didn’t see any, but there are always bugs, right?
Eventually, I started my code base entirely from scratch — I re-wrote and re-checked my math. I re-coded the calculus from the ground up. But the algorithm still failed. However, it failed in exactly the same way it did the first time.
That was a very good indicator that there may be a flaw in the algorithm itself; otherwise I’d have had to introduce the same bug in two separate places in the same way, which was highly unlikely. My team leader was unconvinced, even a bit frustrated with me and my work. He said that if I needed someone else to do the math for me, that I should just ask. There was no shame in admitting that I needed help, after all.
I assured him that I could do it, but that I suspected that my math and code might not be the problem. Perhaps we’d chosen the wrong algorithm for the problem from the beginning. He was staunchly unconvinced. I was, he pointed out, the newest person on the team, and the algorithm to be used was chosen by experts long before I’d arrived.
Eventually, I had to let the matter rest. I couldn’t make progress on the machine learning research on that road, and didn’t know why. I pursued my own ideas. Eventually I hit on an algorithm that DID work, and on the way, I realized my data could clearly explain why the original algorithm could never have succeeded.
When I presented my new tools, and my critique of the old, my team leader only suggested that I had spent too much time spinning my wheels, and that I needed to make better progress. He also thought that the original algorithm still probably ought to work, regardless of my results, which were based on new, untested approaches and he didn’t trust them.
Some months later, I relayed the story to other folks that had worked under him. They were all baffled. Not once, they said, had he ever questioned their results in the way he had done with mine. He’d dive in, engage with them, and figure out what it all meant, and help move forward. He was a "great mentor" and "a brilliant researcher" and he never dismissed them out of hand.
However, the researchers I had spoken with about the team leader were all men.
My takeaway from this past experience is two-fold. First, my team leader looked at me and saw insufficiency and doubt. I had the same track record as my male counterparts, but somewhere along the way, where he read “capable” over them, he read “needs help” over me.
The second, and maybe more important takeaway, was that I had believed him. I looked at my results and his response and thought, “I must be flawed. I must not be as good at this as I thought.” For months, I worked and worked to find what was wrong with what I had done. There was nothing wrong. But I expected myself to fail more easily than my colleagues, just as my team leader did.
There are stories like this everywhere, when you talk to non-masculine identified people in technical fields. When we succeed, it is assumed that we were helped. When we get worn out from the pressure, we are admonished that we might be better off looking for a less-stressful career.
And yet, with the right group of people around me, it can be an amazing field. At Cylance, I’ve been treated like any other member of the team. I get to work with some unbelievably sharp people, and have all the data and algorithms I could ever want! What a difference it can make when your team mates have confidence in you from the start, rather than having to fight through years of bias and doubt before you’ve “proven” yourself to be as valuable as others.
These issues of entrenched, unintentional sexism will be with us for a long time. Our culture isn’t a shining example of one that can change easily for the better. But there are real gains to be had when each of us speaks up in our own circles when we see these biases affecting behavior and actions in real-time around us.
Very few of us will be able to change the entire culture of our time, but nearly all of us can help change the culture in our own spheres. I think that’s where we will start seeing gains. It’s like all the individual yeast cells in a bread dough. Not one can do anywhere near enough work to make the bread rise, but one by one, with their own little bubbles of carbon dioxide, they can contribute to something worth having.
And so can we.