Anchi Li (李桉弛)
AI Researcher & Engineer
Building reliable LLM agents and AI systems for software engineering and human-centered applications.
My work connects research on large language models with robust, production-oriented AI systems.
About Me
I am a Computer Science graduate from Beijing University of Technology, working at the intersection of large language models, AI agents, and software engineering.
My recent work includes coding LLM post-training and evaluation at ByteDance Seed, retrieval-augmented code generation at Concordia University, and human-centered AI research at the Institute of Software, Chinese Academy of Sciences.
My research interests include AI for Software Engineering, LLM Agents & Evaluation, Retrieval-Augmented Generation, and Human-Centered AI.
Research Interests
AI for Software Engineering
Code generation, repository-level understanding, and AI-assisted development workflows.
LLM Agents & Evaluation
Reliable agent execution, task construction, and large-scale evaluation.
Retrieval-Augmented Generation
Retrieval, reranking, and context optimization for knowledge-intensive systems.
Human-Centered AI
AI systems designed around the needs of developers, educators, and other end users.
Recent Highlights
ByteDance Seed
Worked on coding LLM post-training, evaluation, and distributed agent infrastructure.
ACM CHI 2025
Co-authored TutorCraftEase, published at ACM CHI 2025.
Mitacs Globalink
Conducted research on retrieval-augmented code generation at Concordia University.
RAGenius
Independently developed and deployed an end-to-end RAG system with hybrid retrieval and production infrastructure.
Beyond Research
Outside of research and engineering, I enjoy cardistry and photography.
Contact
For research or engineering conversations, email l1anch1@outlook.com. You can also find me on GitHub ↗ and ORCID ↗.