I build LLM agents, MCP tooling, and AI observability — and ship them with production-grade TypeScript, Python, Rust, and .NET.
Vietnam · Remote (UTC+7) · Open to remote roles
- AI agent tooling — tracing, cost profiling, and hallucination detection for LLM apps
- MCP ecosystem — a package manager and a curated hub for Model Context Protocol servers
- Healthcare interoperability — Certified HL7 FHIR Implementer; core contributor to FHIRBridge
|
Self-hosted observability for AI agents — trace tool calls and debug agent runs.
|
Package manager for MCP servers across Claude, VS Code, Cursor, and more.
|
|
AI coding agent in pure Rust — multi-provider, 19 tools, plugins, MCP, sandboxing.
|
Pre-commit proxy that catches hallucinated functions and imports in AI-generated code.
|
|
Curated hub to browse, search, and install 40+ MCP servers.
|
Document Q&A with retrieval-augmented generation, source citations, and streaming.
|
- FHIRBridge — open-source HL7 FHIR R4 toolkit: de-identification, CSV/Excel import, SSRF-hardened self-hosting (25+ merged PRs)
- ai-cost-profiler — token, latency, and cost profiling across LLM providers
- repo-archaeologist — reconstructs the architecture of legacy codebases
- vong-ai-recorder — multilingual speech-to-text desktop app with local Whisper (Rust)
- rt-translator-desktop — real-time meeting translation with Whisper and Ollama
- fhir-validator-web — FHIR validator with Japanese Implementation Guide support
GitHub stats
