Summary
Zhen-huan Yang is an Applied Scientist II with a PhD in Computational and Applied Mathematics and nine years of industry experience building search, NLP, and large-model systems. He has led production search science and conversational search efforts at Amazon and engineered retrieval and ranking pipelines and LLM in-context solutions at Etsy, plus work on text-to-video and recommendation fairness research published at WWW. His recent focus is on efficient LLMs, VLMs and omni-models for edge AI—optimizing distillation- and quantization-aware pretraining and post-training for instruction following, reasoning, and embodied agents. Comfortable spanning theory and production, he combines deep math-driven research in optimization and deep learning theory with hands-on system design for FAISS-backed retrieval, cross-encoder ranking, and LLM-powered buyer experiences. Based in New York, he also explores RL, vision-language models, and quantized deployments, bringing uncommon breadth across model theory, efficiency, and product-facing ML.
9 years of coding experience
4 years of employment as a software developer
Mathematics, Mathematics at University of California, Berkeley
Doctor of Philosophy (PhD), Computational and Applied Mathematics, Doctor of Philosophy (PhD), Computational and Applied Mathematics at University at Albany
BS, Computational and Applied Mathematics, BS, Computational and Applied Mathematics at Sun Yat-sen University