Summary
Haozhu Wang is an AI researcher and Member of Technical Staff based in the San Francisco Bay Area with a decade of experience advancing LLMs, reasoning, alignment, and reinforcement learning. He has driven foundation-model development and post-training techniques across AWS Bedrock, Meta’s GenAI and Superintelligence Labs, and now pushes reasoning frontiers at xAI. His work spans LLM distillation, preference alignment, inference-time scaling, automatic prompt optimization, and model routing—practical research that has been deployed into large-scale GenAI services. A PhD-trained engineer from the University of Michigan, he combines rigorous academic publication and mentorship experience with hands-on product-facing ML systems. Notably, he contributed to Amazon’s Titan models and has published and patented RL and design automation research prior to focusing full-time on large-model reasoning. Colleagues describe him as a researcher who bridges deep theory and production-ready alignment engineering.
10 years of coding experience
7 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Electrical and Computer Engineering, Doctor of Philosophy (Ph.D.) Electrical and Computer Engineering at University of Michigan
visiting undergraduate student Electrical and Electronics Engineering, visiting undergraduate student Electrical and Electronics Engineering at Massachusetts Institute of Technology
Bachelor's degree Electrical Engineering, Bachelor's degree Electrical Engineering at Tianjin University
Bachelor's degree Electrical Engineering, Bachelor's degree Electrical Engineering at Nankai University
Chinese, English