Meet Lisha Li. As a Principal at Amplify Partners, she helps aspiring tech and AI start-ups connect with end-consumer markets and advisors. She’s loved math and theoretical physics since she was a kid, particularly because of her interest in studying subjects that she describes as having a “transcendent” quality.
Learn who her role models are, how she got involved with investing, and why she has a positive outlook for the future of AI.
We interviewed Lisha as part of AI4ALL’s Role Models in AI series, where we feature the perspectives of people working in AI in a variety of ways. Check back here weekly for new interviews.
As told to Nicole Halmi of AI4ALL by Lisha Li; edited by Panchami Bhat
NH: Can you describe what you do as a Principal at Amplify partners? What does a typical day look like for you? What kind of projects do you work on?
LL: Amplify makes very early seed investments and does a little bit of Series A investing as well. We’re looking for exceptional founders who are outstanding members of the field they’re in and are oftentimes first time founders. We typically invest in very technical, machine learning-oriented founders. Once we invest in companies, we try to involve ourselves further by helping out with finding end customers, advisors, and other connections.
I like to talk to people who lead innovation labs in industries that I’m interested in like manufacturing, retail, and agriculture. I learn about what their problems are, then figure out how I can connect them with companies we work with to create a mutually beneficial relationship.
Coming from a technical background, learning about problems and solutions is interesting and is what makes the job fun for me.
How did you decide to get degrees in statistics and math? Were you interested in math at a young age, or did you discover it in college? And how did you come to focus on deep learning and probability?
I went to the University of Toronto to study international relations, but I realized within the first semester that I was craving something else. I wanted to study something that was transcendent in the sense that it had an invariant quality. In my mind, that meant either studying philosophy or looking into math.
With Steven Hawking’s recent passing, I got a little bit reflective about when I first read A Brief History of Time in my childhood. I really enjoyed the idea that theoretical physicists and mathematicians felt that you could get a grasp of reality and answer questions about the origin of the universe through mathematical structures and research. That was a really incredible feeling and that kind of stayed with me. With that in mind, I decided to study pure math because if I ever wanted to go into theoretical physics I would be able to understand it.
Deep learning and probability was a very natural transition given my background in graph theory. There was a lot of exciting work happening in deep learning, and I wanted to contribute both empirically and with some theoretical foundations. Before I interned at Pinterest, I thought there was a 50–50 chance of me pursuing academics or going into industry. After the internship, I knew 100% I would go into industry work.
How did you decide to move into investing?
Initially, investing wasn’t even a career that I was considering. I met my husband in the first year of grad school, but he later dropped out of the math PhD and became a data scientist at Square, then he ended up going to the Stanford Graduate School of Business (GSB). The GSB was where I met people in the investing world, and I realized the work they do is pretty interesting.
I hadn’t realized that early stage investors spend a lot of time trying to translate ideas into real business impact. That caught my interest, so I started talking to more venture capitalists.
Then, I ended up finding Amplify. They were special to me because they’re technically focused and also enterprise focused. I also felt like I could learn a lot from the partners. They valued the skills I was coming in with, and were open to training me in areas where I didn’t have experience, which allowed me to come in at the mid-level. I took a chance and it turned out great.
You grew up largely in Toronto, Canada. Do you find that there are any differences in the perception of and discussion about AI between the Bay Area and Toronto?
Toronto holds a special place in the recent AI revival, in particular with Geoff Hinton and his work on neural networks. Toronto is trying to make use of the talents that they’ve grown within the community. Even so, it’s difficult to recreate Silicon Valley’s culture, even in cities that are closer to Silicon Valley’s maturity of tech talent, such as Seattle and New York.
In Silicon Valley, attempting to pursue a start-up is treated as a more serious endeavor, even if you don’t do well. I go back to Toronto and try to lend any experience and insights that I may have. I mentor start-up companies in Toronto through a great program called the Creative Destruction Lab.
People are just as concerned with ethics in Toronto and are becoming increasingly alarmed about ethical considerations.
What are some of the important things people should be doing now to create a positive, inclusive, and ethical future for AI?
Because we rely on AI technology for predictions and to inform our behavior, being aware of the bias we as people hold is a key step into knowing how to design tech properly.
From my experience, people are on board with making tech education accessible, but they’re not sure how to execute on that. I’ve seen a lot of more great grassroots mechanisms such as Fast AI, founded by Rachel Thomas and Jeremy Howard. They do a great job ramping software engineers up in deep learning. They’re also working hard to make that knowledge accessible through things like their diversity fellowship.
When we educate non-technical individuals about the impacts of AI, we need to help them be imaginative about AI. It’s not just about killer robots or random results. There are a lot of positive things that can come out of AI, as long as we know how it works. In my talk Differentiable Programming as a Framework for Machine Intelligence, I tried to give a framework to understand these advancements.
Who were your role models growing up? Do you have any role models now?
I credit my parents with more of a laissez-faire upbringing — which is very strange for a Chinese immigrant household, but I reacted positively to that parenting style. My parents were academics in China and uprooted themselves to give me a better life. I feel a sense of wanting to not waste that effort. My mother especially is a very hard worker and that always inspires me to work even harder than I am.
In my childhood, I remember watching a historical drama series in Chinese about the only female empress in Chinese history, Wu Zeitan during the Tang dynasty. The series highlighted her contributions to establishing a lot of meritocratic systems in China’s education system.
Even as a kid I realized women weren’t in a lot of positions of power, so I found the story inspiring and it helped me see that women have succeeded in much harder times.
Between this story and my mom, I didn’t have a limiting outlook on what I can accomplish, gender-wise.
What advice do you have for young people who are interested in AI who might just be starting their career or academic journeys?
I think there’s interesting theoretical work to be done in AI. Don’t shy away from taking the math courses that might help inform this. I think there are a lot of really fruitful connections with theoretical physics and statistical mechanics.
Since this field is so young, having collaborations with people who work in other areas that have nothing to do with deep learning is useful. Studying something that you enjoy is a good way to introduce new ideas and collaborations into the field as well.
My outlook is very positive, and I think it’s a great time to be getting into this field.
About Lisha Li
Lisha Li invests in technical founders solving ambitious problems. From compute substrates to the creative process, medicine to manufacturing, she is excited to be investing at a time when machine intelligence and data-driven methods have such incredible potential for impact. Investments she has been involved with include Embodied Intelligence and Primer. Lisha completed her PhD at UC Berkeley focusing on deep learning and probability. While at Berkeley she also did statistical consulting, advising on methods and analysis for experimentation and interpretation, and interned as a data scientist at Pinterest and Stitch Fix. She was the lecturer of discrete mathematics, as well as the graduate instructor for probability and computer science theory. She is @lishali88 on Twitter and @lapis.lazuli.8 on Medium.
Follow along with AI4ALL’s Role Models in AI series on Twitter and Facebook at #rolemodelsinAI. We’ll be publishing a new interview with an AI expert every week this winter. The experts we feature are working in AI in a variety of roles and have taken a variety of paths to get there. They bring to life the importance of including a diversity of voices in the development and use of AI.