An anesthesiologist building practical Medical AI, clinical agent workflows, and open-source research tools.
I work at the intersection of perioperative medicine, life-science machine learning, and AI agents. My focus is turning real clinical questions into tools that can be inspected, tested, reused, and improved in public.
Personal homepage · GitHub projects
- Medical AI systems for clinical reasoning, anesthesia safety, health-report interpretation, and emergency workflows.
- AI agent workflows for research assistance, clinical decision support, and repeatable knowledge work.
- Open learning resources that make agentic engineering and medical AI easier for beginners to understand.
- Life-science ML experiments inspired by Kaggle competitions, bioinformatics, and structured clinical prediction.
- Silver Medal · Stanford RNA 3D Folding
Rank 55 / 1,516 teams, Top 3.6%. - Silver Medal · CIBMTR Post-HCT Survival Equity
Rank 83 / 3,325 teams, Top 2.5%. - Bronze Medal · CZII CryoET Object Identification
Rank 51 / 931 teams, Top 5.5%.
See all Kaggle competition results: kaggle.com/lixinyin/competitions
| Project | Why it matters |
|---|---|
| periop-anesthesia-agent | Doctor-reviewed clinical AI agent template for perioperative anesthesia assessment. |
| clinical-ai-agent-skills | Evidence-first agent rules for clinicians using Codex, Claude Code, Cursor, Copilot, and Obsidian in clinical research workflows. |
| multi-agent-collaboration-learning | A multilingual Medical AI Agent field guide for healthcare learners and AI builders. |
| openai-agents-medical-research-guide | A safety-first course for building medical research assistants with the OpenAI Agents SDK. |
| agent-token-budget | A visual guide to token budgets and safety boundaries for AI coding agents. |
| med-dataset-platform | Tooling for medical AI dataset generation, management, and export. |
- Merged: academic-research-skills #236 — added a clinical citation verification checklist for detecting unsupported medical claims and overconfident deployment language.
- Merged: academic-research-skills #235 — fixed a stale EMNLP disclosure policy URL.
- Open: academic-research-skills #240 — proposed a clinical epistemic-status example for silent upgrade detection.
- Open: awesome-claude-skills #904 — submitted Clinical AI Agent Skills as a healthcare-adjacent Claude Skill resource.
- Open: andrej-karpathy-skills #154 — added a domain-specific safety customization example.
I am especially interested in:
- Clinical AI agents that are transparent enough for doctors to audit.
- Medical research copilots that support, rather than replace, human judgment.
- Reliable open-source workflows for clinicians learning AI engineering.
- Small, reviewable contributions to serious AI, healthcare, and research tooling projects.
If you work on medical AI, agent frameworks, life-science ML, or clinical research tools, I am happy to discuss practical workflows, documentation, reproducible examples, and small open-source contributions.
我是麻醉医生，也在持续把一线临床问题转化为可验证、可复盘、可开源的医疗 AI 与 Agent 工具。