Jason Li

Hardened key derivation in multi-party computation

US patent 18/352,914

Developed a method for achieving hardened derivation efficiently when running multi-party computation algorithms such as GG20, FROST, Lindell17, etc., while preserving key refresh and the recoverability of derived child keys from root keys and the chaincode. This method is used in enterprise-grade crypto custody systems to offer enhanced isolation and greater security guarantees.

Wavelet: Efficient DNN Training with Tick-Tock Scheduling

Proceedings of Machine Learning and Systems · Mar 15, 2021

Devised Tick-Tock scheduling system to accelerate distributed GPU training. Results with DNN models showed a 6.7x training time reduction.

Risk Analysis of Cryptocurrency as an Alternative Asset Class

Applied Quantitative Finance. 3rd ed., Springer-Verlag Berlin Heidelberg. ISBN 978-3-662-54486-0

Analyzed the risks of cryptocurrencies as an alternative investment. Specifically, examine the wealth distribution of cryptocurrency, evaluate its corresponding effects on the market, and analyze other risk factors leading to the demise of altcoins. The paper concludes that the closer the right tail of the wealth distribution approaches the Power-Law model, the more stable the market will be. This result is valuable for investors in making decisions when investing in cryptocurrencies.

CS161 Computer Security

Led the design and implementation of Projects – SQL Injection, CSRF, Reflected XSS, Code Injection, Clickjacking.

CS162 Operating System

Teaching discussion sections, conducting office hours, contributing to exam questions, and managing cheating detection for projects.