Research
I study how markets tell truthful information from false, and how that shapes long-term value, especially now that AI agents take part in markets.
The question
Discerning truth from false information is, I believe, the key to identifying long-term, sustainable value, both for investors judging companies and for platforms where AI agents now make claims at scale. I want to understand which market mechanisms reward honest participants.
- When AI agents make claims in a market, does staking money on accuracy reduce false claims more than reputation systems do?
- Under staking, do honest sellers earn more over time, and how quickly does the market learn who is honest?
- In financial markets, how quickly do prices correct after false corporate claims are exposed, and what speeds that up?
Current work
Market Design Interventions for Safer Agentic AI
Research Intern · Prof. Marshall Van Alstyne · Boston University · since Sep 2026
An experiment in which people and LLM-powered agents trade as buyers and sellers, testing whether staking on advertised claims makes market information more reliable. I proposed an asynchronous, continuous-time market design and independently built a full-stack oTree prototype of it. I’m now developing LLM agent integration and data capture in the team’s Empirica implementation.
In progress: buy-side research on SK Hynix and TSMC, examining AI-driven demand across the chip supply chain, to be published on chenquant.com.
Background
I bring training in physics (BS, Sun Yat-sen University), economics (M.Phil., Bocconi University, on a full PhD scholarship with first-year PhD coursework completed) and management (Lingnan–MIT Sloan MBA, thesis on gold price movements), plus a decade of practice as a founder and fund manager. I’m completing an MS in Software Development at Boston University.
GRE Quantitative 170/170 · CITI Human Subjects Protection Training · CFA Level 3 candidate
