On the Record with Sarah Nagy, CEO of Seek AI
Hometown: I grew up in the Bay Area, but I moved to LA when I was 11. I got to experience both Northern CA and Southern CA!
Hobbies: Music has been a hobby of mine ever since I was about 5 years old and started playing classical piano. I played piano for 12 years and then branched out to produce rock albums, produced electronic music, and DJ’d. I’ve also surfed here and there over the years but never in NYC--you never know, that might become my new hobby.
3 words to describe you: ambitious, creative, builder
Favorite hotspot in NY: Brandy Library (always a classic)
What were you like as a child and how do you think that shaped who you are today?
When I was a kid, my parents didn’t really care what I was going to do with my life when I grew up as long as I was happy, and I feel very fortunate to have grown up with that parenting philosophy. Where I grew up—in the Silicon Valley and then the LA suburbs— parents were always pushing their kids, and I didn’t have that. That said, I was ambitious, even as a kid, and in a competitive sort of way. I would see my friends doing sports or piano lessons or honors classes, and I would say to myself “well, I can do that, too.” But my parents were very “go with the flow” and supportive people who helped to enroll me in different activities. I think they ultimately wanted me to be an artist, though. My mom still brags about a project I did when I was in early elementary school, reminding me that I drew a leaf under another leaf when I was drawing a tree. She claims that was advanced spatial reasoning for my age!
Since you ended up becoming an engineer, rather than an artist, what do your parents think now?
I actually haven’t shown them the latest version of Seek! I showed them an early UI before I incorporated it, but at that point it was more a pre-Seek product than our actual, current product.
My parents are supportive but they are somewhat risk averse so I think they are excited for me, but my dad gives me very practical advice. There's not a lot of fluff in those conversations. It’s more, “do you have customers?” “how well does it work?”, etc… He was a hardware engineer in Silicon Valley, though, and knows quite a bit about deep tech.
How did you become a data scientist?
I started out as a quant. I was a double major in college in astrophysics and business economics, and while I was always interested in finance, I didn’t really see myself as a fit in investment banking. I also didn’t like accounting math as much as, say, partial differential equations for astrophysics. So, when I learned about a field of finance where I could solve math problems, all that time, *that’s* what felt like the perfect fit for me. After getting my undergrad at UCLA, I was accepted into Princeton’s Masters of Finance program. I was very lucky to go there and learn from the best about quantitative finance. I started out as a quant at a trading firm called ITG where I was in the weeds solving trading execution problems. Around 2013/14 I started to hear about machine learning, and I could already tell that it was going to be a huge field. ML was starting to get used in the financial world, and I could see it adding more value than traditional techniques. Back then data science and ML were intertwined, so I moved into data science, but what I really wanted to do was ML on big datasets, which was a totally new field. That’s how I got into the startup world because there weren’t a lot of ML opportunities in the big financial firms back then. At Predata, I got to work with Twitter, Youtube and Wikipedia data, which was really interesting. I have been a data scientist from then on, and the field has grown a lot.
What ultimately led you to start Seek AI?
I have been really excited about large language models writing code from the moment GPT-3 came out. I’d been reading Hacker News for a long time, and when I saw it pop up on there, I thought it was a big innovation. I was more excited about this than I had been about any other tech in a long time–probably since I saw VR for the first time in 2013. As a data scientist who writes code to query data all day, I could immediately see the potential in seeing someone ask GPT-3 to generate code to pull up a list of countries by GDP, for example. I didn’t know exactly what exactly I was going to build, I just knew I had to make a bet on this technology. It inspired me to focus on learning more about these types of models, which I did for several months. Once I had really learned about the models, that’s when I realized I could solve a pain point I encountered everywhere I had worked as a quant or as a data scientist.
The pain point in question was business people not having the right tools to answer their own data questions. As a data scientist, I would frequently work on problems that required a lot of focus, but I was often interrupted by the business side people who had questions about the data, forcing me to stop what I was doing. The process seemed archaic and inefficient. I realized that if I focus this new technology on solving the problem, it could be category defining.
How did your founding team come together?
Interestingly enough, Sarah Smith was actually our first advisor. We were working together at a company called Edison building out the NY office—me with the data science team and her with the sales team. The NY business ultimately contributed quite a bit to Edison’s acquisition, with the acquirer (YipitData) becoming a unicorn after the acquisition. She was on our advisory board for about 6 months, at which point I had validated the product and raised the pre-seed round, and it was time for me to bring on some cofounders. Sarah and I agreed it made sense for her to join. David Lee, our Head of Engineering, joined around the same time as Sarah, and I was introduced to him through a mutual colleague. You can’t learn everything about someone the first time you meet them, but as I've gotten to know and work with David, it’s even more clear that he is the perfect person to be our Head of Engineering. He had actually worked with very relevant technology at his prior company, DataFox. He has experience in data engineering, MLOps, human in the loop apps, and as well as other relevant applications. I met Raz Besaleli through my efforts to bring on someone to focus on the NLP algorithms, and she is now our NLP Research Scientist. What stood out to me about her is that she comes from not just a deep learning background but also a linguistics background. Of course, she also knows a lot about deep learning, but that knowledge of linguistics has been really helpful because if you don’t know much about how language is structured, it could definitely be a disadvantage.
What excites you the most about the work you’re doing at Seek AI? What is Seek’s “superpower”?
What excites me the most is that by using AI, we are building products that can understand questions and generate knowledge by querying databases—that has a lot of room to go in terms of meaning. That’s why our slogan is “what matters” because there’s a lot more to Seek than just querying databases. In 10…20 years from now, Seek might be able to help researchers look at huge datasets that no human could totally analyze on their own and help discover things that we couldn’t otherwise discover without AI. The idea of that really excites me, and I can’t think of a better way to spend my time.
In the shorter term, though, it’s really cool to see people’s eyes light up when they see what NLP can do. A lot of people don’t know NLP can write code.
Have you ever needed to change or pivot the business based on product feedback, product market fit, or product building?
The core of our product is to make it easier for organizations—both tech and non tech teams—to work with data, and we have not had to pivot away from that. We have learned new features, though, and discovered new ways to integrate NLP into the product that I hadn’t thought of before.
Can you talk a bit about the company culture at Seek?
There’s a few principles I try to adhere to at Seek. Firstly, a big part of our company culture is to start from “first principles” vs alternatives, such as using analogies or copying what others are doing without really understanding *why* you’re copying them. Coming from an investment background, I’ve learned that you have to think for yourself—come up with a strategy to make investments that others don’t know they should be utilizing. So this is what I try to stress at Seek. We are doing what we are doing and not trying to emulate anyone else.
Another leadership philosophy I adhere to is making sure the team is aligned towards specific product goals. I don’t want to distract the team with lots of different goals we may hear from customer feedback. We’ve had customers suggest features that are good ideas, but we can’t work on every little thing, so we have to be focused on our hypothesis direction.
I also believe that done is better than perfect. There’s a quote from Sun Tzu, “I have heard that haste can be folly, but have never seen delay that was wise.” Obviously I’m not saying to be negligent, but if you’re building a prototype, and you build it fast, then you can at least get feedback which is extremely valuable. From what I've heard and read, that’s how the best startups win.
Lastly, related to the pain point which was the impetus for Seek, I live by the converse principle as well: “never make the business people work with the data”. They don’t want to touch the data. They just want the answers so don’t burden them with any of the querying of the data. That’s a product principle but I think about it every day.
Overall, I’m really happy that none of us get defensive when it comes to differences about work. We’re all honest about what’s working and what’s not, and I think that’s an essential part of any startup culture.
How did you get connected with Differential Ventures and what does the relationship mean to you?
Once again, it actually started with Sarah Smith! Sarah connected me to Jeremy Baksht who connected me to Melissa from DVP. Sarah and Jeremy knew each other from the NY finance community, and when I talked with Jeremy he got really excited about Seek and wanted to help me.
DVP has been amazing to partner with. Firstly, in terms of the background–you don’t meet a lot of super quants (David) who worked at Renaissance and were key parts of Renaissance’s success during its heyday! He also has an NLP background, which is a big plus. It’s an honor to work with David—honestly I was super excited just to meet him. Something in particular that stood out to me about David is that he wants to hear all the details about how Seek’s product works—he wanted to learn more about the weeds, which I would not say is typical in the VC-startup relationship. Nick also comes from an NLP background, and it’s really awesome to work with a team like this and be able to bounce both product and GTM ideas off of them. DVP is truly thinking about how Seek can grow in the future—about how to truly make us a category defining category.
Who has been your biggest inspiration as an entrepreneur?
I probably have a roster of about 500 different entrepreneurs that I follow. I love podcasts and YouTube videos about entrepreneurs. Jim Shinn, CEO of Predata, and Michael Berner, CEO of Edison, are mentors to me that I had the privilege of working alongside before starting Seek. In the public eye, I look up to Bob Muglia, the former CEO of Snowflake which grew into the biggest software IPO in history at the time. I learned a lot from Snowflake's success. I also recently listened to a podcast with Cindy Eckert and think she is an entrepreneur who really stands out. She’s not in my space, but her story is really cool.
—
Sarah started Seek back in September 2021, and comes from a quant/data science background. Sarah started her career as an astrophysics researcher and then became a quant doing algorithmic trading at ITG. She then led the quant efforts at two startups (Predata and Edison, both acquired by unicorns) as well as the consumer data team at Citadel’s Ashler Capital. Sarah has B.A./B.S. degrees in Astrophysics and Business Economics from UCLA, and an MFin degree from Princeton.
Related News
