Summary
Zhiyuan Peng is a Machine Learning Engineer and Ph.D. candidate in Computer Science based in San Jose with eight years of experience building production-grade NLP and search systems that bridge research and business impact. He led the development and deployment of an entity-aware multi-task query understanding model at Walmart that cut GPU usage by 75%, improved P99 latency by 18.6%, and raised GMV by 0.51%, and has published related work at KDD and SIGIR. His background spans e-commerce search, dense retrieval, and reranking across companies like Walmart, ByteDance/TikTok, and Navan, translating LLM and IR research into scalable ML services. Comfortable across the stack—from model research to low-latency productionization—he also has hands-on experience with annotation, inference optimization, and cloud deployment. An uncommon blend of rigorous academic training and measurable product outcomes makes him effective at turning advanced models into live revenue-driving features.
8 years of coding experience
5 years of employment as a software developer
Bachelor's degree Electrical Electronics and Communications Engineering, Bachelor's degree Electrical Electronics and Communications Engineering at Beijing Jiaotong University
Ph.D. Computer Science, Ph.D. Computer Science at Santa Clara University
Master's degree Electrical Electronics and Communications Engineering, Master's degree Electrical Electronics and Communications Engineering at Beijing Institute of Technology