Summary
Zhi Lee is an LLM algorithm engineer based in Beijing with a strong research pedigree from Tsinghua University and hands-on startup experience in AI productization. He designs robust retrieval-augmented generation systems—most notably ReaRAG—and explores neuro-symbolic approaches to knowledge-based QA, blending structured knowledge and symbolic logic with generative models. His background spans semi-supervised learning and graph neural networks from Shanghai Jiao Tong University and production work on face anti-spoofing and digital avatars, where he led end-to-end R&D to meet ISO presentation-attack standards. Though early in his career, Zhi combines rigorous academic training with practical system-building, often stress-testing models with adversarial distractors and long-context baselines. He also brings a concise, principled research style evident in evaluation-focused projects and publicly shared work on his personal site.
1 year of coding experience
3 years of employment as a software developer
SMK Cochrane
GCE A-Levels (Science), GCE A-Levels (Science) at Tunku Abdul Rahman University College
Master in Advanced Computing, Master in Advanced Computing at Tsinghua University
Bachelor of Engineering in Automation, Bachelor of Engineering in Automation at Shanghai Jiao Tong University
Chinese, English, Malay