Hello, I'm Edward Yu. I'm a builder at heart, and the founder of Variational Research, a technology firm building peer-to-peer decentralized derivatives. I'm interested in Bayesian statistics, machine learning/AI, quantitative trading, and decentralized applications.

The cities in which I spend the most time these days are (in no particular order): New York, NY; Berkeley, CA; Taipei, Taiwan.

Contact me at [email protected]

Professional stuff

• 2021-Present: Variational Research - I founded a protocol to enable bilateral derivatives on any underlying time series. We’re a small team backed by venture capital. News: $10m seed raise by Bain Capital Crypto, Peak XV, Coinbase Ventures, Dragonfly, others.
• 2019-2021: Genesis Trading - I was Head of Quantitative Research for one of the largest trading firms in crypto, doing market making on linear and options markets.
• 2017-2019: Qu Capital - I founded a stat arb hedge fund trading crypto. We raised money from various institutional investors and eventually sold it to Digital Currency Group (and its subsidiary, Genesis Trading).
• 2014-2017: I studied Applied Mathematics at Columbia University, graduating with honors with a Bachelor of Science.

Writing

• (2025-07-11) How to price: pump.fun ICO

• (2025-07-05) A collection of quotes implying hedging is bad

• (2025-05-21) How to Price: Will HYPE Overtake SOL By End of Year 2026?

• (2025-05-16) The Fairest Casino

• (2025-05-06) Programmable Derivatives: a Vision for Variational’s Future

• (2025-04-27) Hacking Your Greeks: A Guide for Retail Options Traders

• (2025-04-25) I want to start writing more

• (2023-09-04) Lessons from trading

• (2020-10-11) Nassim Taleb calls Nate Silver totally clueless about probability: who is right about election forecasting?

Research

• Bayesian Neural Networks with Soft Evidence. Presented at the 2021 International Conference on Machine Learning workshop on Uncertainty & Robustness in Deep Learning
• Tree-based Pipeline Optimization Tool. In 2021, I contributed some features to this open source library for AutoML.
• A Bayesian Ensemble for Unsupervised Anomaly Detection. 2016 arxiv preprint, research was done (and implemented into production!) while I was at Facebook

Investing

My investing philosophy: many diversified alphas and a healthy dose of leverage.

I’m an angel investor in ~20 startups. I prefer seed stage or series A.