Zhen Li is a Staff AI Engineer based in Beijing with six years of experience building high-impact ML and systems products at the intersection of research and production. He led Replit’s AI Agent as creator and lead engineer, helping the product scale to $100M ARR within six months, and previously co-founded a VC-backed AI startup focused on workflow automation and code generation. His background includes productionizing large-scale user models at Google and recommendation/ads systems at ByteDance, plus performance-focused C++ contributions to the widely used Velox execution engine (notably HashTable and join optimizations). Zhen combines deep systems engineering—file-handle caching, prefetching, and decimal support—with applied ML, making him effective at squeezing latency and cost out of real-world AI services.
A composable and fully extensible C++ execution engine library for data management systems.
Role in this project:
Back-end Developer
Contributions:652 reviews, 102 PRs, 998 comments in 1 year 11 months
Contributions summary:Zhen primarily focused on optimizing the performance of the Velox execution engine, specifically within the `HashTable` component. Their work involved optimizing aggregation by avoiding redundant computations of normalized keys in hash tables. They also improved the performance of join probes, adding prefetch commands for tags and increasing the group size in join operations. Additionally, the user documented and implemented file handle caching and added support for decimal data types within various functions, increasing the functionality of the engine.
Contributions:352 pushes, 100 branches in 2 years 5 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.