Summary
Corey Hu is a Machine Learning Engineer specializing in reinforcement learning and post-training work on large language models, currently contributing to Meta Superintelligence Labs. Over two years in ML roles, he helped build LLM observability and evaluation tooling (TruLens) through TruEra and Snowflake, and earlier delivered production deep-learning solutions at NVIDIA that achieved dramatic runtime and accuracy improvements for chip design tasks. His background spans applied research in computer vision, optimization, and model debugging—from published work on transformer-based EDA tools to log-file root-cause classification—demonstrating a blend of research rigor and production impact. Based in California and Berkeley-educated in computer science, he moves models from prototype to scalable deployment and retains a penchant for solver-driven, convex optimization approaches beneath modern deep learning. An often-overlooked strength is his cross-domain track record: applying transformer architectures not just to NLP but to hardware design and operational observability.
1 year of coding experience
7 years of employment as a software developer
Bachelor’s Degree Computer Science, Bachelor’s Degree Computer Science at University of California, Berkeley
English, Chinese