Accepting some of the truths of being a woman in tech
With the upcoming issue focusing on Women in Tech and Female Founders, we decided to speak to a few females in the industry on their journey and experiences. Starting with Franziska Kirschner, Research lead at Tractable. "When I was a child, I wanted to become a dog when I grew up. I genuinely believed that science would progress far enough that I could change species by the time I reached adulthood. Unfortunately, the world’s scientists had other ideas and their endeavours focused elsewhere, such as on making computers reason like humans.
The result: well, my dream life of sleeping on the sofa seems unlikely. Conversely - and happily - other breakthroughs have occurred; for example, deep learning now allows AI to classify images to the same accuracy as humans. Human food tastes better than dog food anyway.
Today, I’m a research lead at Tractable - Startups Magazine’s UK startup of the year. Tractable uses that good news on the image classification and applies it to accident and disaster recovery.
To describe it briefly: you crash your car, you take some photos of it, the AI looks at those photos and tells you how damaged the car is, you get your car repaired, your insurance pays for it, and life’s back to normal. When done by humans, the process from crash to cash takes weeks. When done by AI, it takes hours. It’s cool.
Before that, I was a physicist - I studied for seven years at Oxford, had published in Nature, and felt I kind of knew what I was doing. It was a great deal of fun, but after a while, I fancied trying something more applied. In what I feel was a momentary lapse in judgement on their side, Tractable offered me an internship. In the space of two years, I went from Deep Learning Research Intern to Deep Learning Researcher, to Senior Deep Learning Researcher, to Research Lead (which thankfully is much easier to type).
It’s a very fast progression, and comes served with a hearty portion of impostor syndrome on the side. I think that’s one of the awesome things about working at a startup, though - how many people in their mid-twenties can say they lead a research group?
My team takes deep learning research from the lab to the real world, bringing benefits to real people and their cars. It really works, but it’s not easy. Throughout our work, we’ve had to learn a lot of lessons quickly - some the easy way; most the hard way. Let me spare you a little of that trouble by sharing the top five hard truths I’ve come to accept:
- You’ll never truly feel like you know what you’re doing
When I started at Tractable, my knowledge of machine learning came entirely from Youtube videos. I felt very out of my depth. Whenever I started getting comfortable, I took on a trickier project - and got relegated right back to the deep end. It’s scary, but it also meant I was pushing the boundaries of my own ability, and learning something new - which is what I wanted in the first place.
- Being a woman working in tech is fine (if you’re working with the right people)
Everyone has heard a ton of inspiring stories on being a woman in tech - as well as a heap of bad experiences and stereotypes. When I was job hunting at the end of my PhD, I really wanted to make sure I ended up working somewhere where I didn’t feel different because I was female. I haven’t worked out a magic formula for this yet, but gut feeling seems a good starting point. When I joined Tractable, there weren’t any other women in tech on the team and I didn’t have much to go on, but my interviewers were honest and genuine and I never felt I was being employed or interviewed to tick a box. If you work in the right place, you don’t need to feel conscious of the fact you’re the only woman in the room.
- Believe in yourself!
I work in a fast-growing startup, and every day there are many opportunities for lively debate. Initially I found this intimidating - it seemed everyone else knew a lot more than I did, and in discussions I tended to be quite ‘British’, so I didn’t always speak up. But soon, I realised that sometimes I did know a lot and have relevant experience, so I made a conscious effort to contribute, which really meant I had to wrestle with my self-doubt and believe in myself and my expertise. It’s important to say to yourself: the work I’ve done and my past experience is unique; I can bring novel ideas and perspectives to the table; and no matter how important or experienced everyone else is, they will never quite have that same angle on a problem that I do, which is in itself valuable.
- Selling your work is an art
I do machine learning because it’s cool. That works well for encouraging kids to follow my career path, but not necessarily for encouraging customers to buy my AI. When talking about my work, it’s important to think “what does the audience want?” rather than “what do I want to tell the audience about it?” It makes it a lot easier to excite an audience if you have their interests at the centre of your work. I take this approach not only to selling what I do, but to planning it too - decide what impact you want on your end-user first, focus on what they want, then work out the steps to get there.
For me, life is not about optimising metrics, accumulating money, or burning GPU hours like it’s going out of fashion. It’s about doing new and sweet stuff with powerful tech, and deploying it across the world to help a whole range of people. This mission is at the core of everything my team does, and I try hard to make sure we enjoy the things we do to achieve it. It’s important to remember: we exist now at a pivotal moment in humanity - the watershed where AI turns from an academic curiosity into a paradigm-shifting technology. AI will significantly change and improve our lives, on par with the invention of electricity and the Internet. Being a part of that change is very cool."