From physics to data science
In hindsight, things often look a lot more linear than they seemed at the time. Same goes for my career trajectory. Motivated primarily by curiosity, I pursued an undergraduate degree in physics and mathematics, before getting a PhD in statistical and computational physics. After graduating at the start of the pandemic, I was lucky enough to get a prestigious fellowship to pursue my own research project at NIST. After a couple of years doing so, I realized I wanted to try pursuing other paths, due to both a frustration with academia and an interest in other fields.
Transitioning into data science in 2022, I’ve ended up working at the intersection of web3, cryptography, game theory, and network science, bringing rigorous quantitative research to practical applications in multiple fields. I am excited to be somewhere with opportunities for me to learn and grow. I love coming up with quantitative solutions to tough problems and figuring out how complex systems work, but I’m realizing that I’ve just only started to explore the countless domains where that is possible.
For more details on my career path, see my LinkedIn page
Data enablement and value attribution
The world of decentralized identity and verifiable credentials
As of March 2023, I joined a new start-up, Valence. We are building a decentralized system for identity/data verification, data exchange, and data valuation. We plan to incorporate cryptographic verification and trust-less, decentralized networks to ensure data remains private and secure while retaining the ability to verify its source. As part of the research team, I get to work on some really exciting new ideas in the worlds of cryptography, machine learning, game theory, and network science!
Recently, a lot of my work has been focused on value attribution: the idea of identifying the components of a system that contributed the most towards the systemic outcome (i.e. the most valuable parts). This research has taken me down some interesting roads in game theory and network science. I’m particularly excited about the concept of Shapley values (see my technical notes on the subject here). In short, this is a game theoretic tool to calculate the contribution of each player in a cooperative game. First introduced in 1951, this idea has gained a lot of traction in recent years as a way to explain results of AI models, which can often be a black box. Intriguingly, there are ways to apply this concept to the study of networks and network centrality (notes here).
Cryptoeconomics
My first data science role was at a crypto startup that built one of the first decentralized exchanges (DEX) in the Cosmos ecosystem. Working on the cryptoeconomic team, I helped develop a python-based simulation framework which can replicate the economic state of the DEX and any agents that interact with it. In my day-to-day I alternated between testing model agreement with live data and researching the potential impacts of new policies or newly developed features. While short-lived, it was a really interesting experience and helped me realize how the simulation skills I’d developed, while studying atoms and particles, could really be applied to a multitude of topics.
Ongoing learning
I consider myself a lifelong student. Navigating a transition from particles and forces to market makers, game theory, and value attribution has been tricky but rewarding. Above all, it has required an ability to quickly learn about new topics. At my previous role, I quickly learned about the economic principles behind a DeFi system with automated market makers. More recently, I’ve delved into cryptography and network science: trying to better understand how we can build secure, decentralized systems at scale. From zero-knowledge proofs to decentralized identity, there is a lot to learn. Nowadays, I’m very excited about game theory and have started reading Maschler et al’s book on Game Theory and Algaba et al’s Handbook of the Shapley Value.
Always open to suggestions for interesting topics or resources to explore.