I am He Dong, an M.S. student at the Beijing Institute of Technology and a former Agent Evaluation Infrastructure intern at ByteDance. I build AI systems that can do useful engineering work while remaining observable, reviewable, and safe to operate.
My work sits at the intersection of LLM agents, AI for code, developer tooling, and distributed reinforcement learning. At ByteDance, I worked on AI-assisted engineering workflows that reduced an MR defect-tracing and code-review cycle from 30 minutes to 3 minutes.
observe → reproduce → locate → patch → verify
evidence at every boundary · humans retain the final decision
I contribute production fixes to widely used developer tooling, with an emphasis on lifecycle correctness, failure containment, and regression-proof delivery. Three of my pull requests have been reviewed and merged into Everything Claude Code (ECC), an agent-harness ecosystem with 250k+ GitHub stars.
| Merged upstream contribution | Engineering impact |
|---|---|
| ECC #2992 · Manual Claude hook lifecycle | Built ownership-aware, idempotent hook registration across install, upgrade, doctor, repair, --no-hooks, and uninstall. Added fail-closed validation, serialized writers, atomic private writes, and 177 focused tests across a 24-file change. |
| ECC #2971 · Loss-safe observer archival | Required an explicit completion sentinel from the current analysis run, preserved observations on process and semantic failures, and cleaned up subprocesses and temporary artifacts after timeout or interruption. |
| ECC #2194 · Unicode-safe CI | Replaced pictographic ratings with meaning-preserving ASCII scores across source and localized docs, unblocking the repository's Unicode-safety gate with a minimal, auditable patch. |
Review all merged ECC contributions ↗
TypeScript AgentTeams MCP PostgreSQL
An evidence-first AgentTeam for repository maintenance. It moves an issue or failed CI run toward a verified pull request while preserving decisions, approvals, tool calls, and rollback points in a tamper-evident execution chain. Interactive demo ↗ · Verified PR |
Rust tree-sitter AST CLI
A safe, example-driven codemod engine. Give it before-and-after snippets and it infers a structural transformation, scans a repository, previews the diff, and generates a rollback path before writing. Quick start ↗ |
Python PyTorch Ray Rust
A distributed Actor–Learner training system for MuJoCo with parallel rollout collection, PPO optimization, multi-GPU experiments, a Rust replay buffer, and real-time experiment observability. See the system ↗ |
Python Click Cross-platform
A local-first CLI that makes Conda environment backup, restore, clone, offline packaging, cleanup, and diagnosis safer and more repeatable across Windows, macOS, and Linux. Project site ↗ · PyPI |
Python HTML PPTX AI Skill
An AI presentation skill that produces animated, zero-dependency HTML decks and editable PowerPoint files through a deliberate theme system designed to avoid generic AI aesthetics. |
TypeScript Agent System Testing
An agent-based system that turns product requirement documents into structured test-case suites, connecting product intent with repeatable quality workflows. |
| Principle | What it means in practice |
|---|---|
| Evidence before confidence | Reproduce the failure, capture the trace, and make every conclusion inspectable. |
| Safe autonomy | Give agents useful tools, explicit boundaries, human approval gates, and reversible actions. |
| Systems over demos | Build typed contracts, tests, observability, and deployment paths around the model. |
| Performance with a baseline | Measure against a frozen reference and promote changes only when regressions are understood. |
I am currently exploring reliable agent infrastructure, repository-scale code transformation, evaluation systems, and scalable reinforcement-learning workflows. I am always interested in thoughtful collaborations around these areas.