Building intelligent systems with AI/LLM integration, scalable backends, and production-grade code.
📍 Singapore | 💼 Open to remote opportunities | 📧 64lamei@gmail.com
I'm a backend and AI systems engineer with hands-on experience in production-grade AI/LLM platforms. I've contributed to and deeply analyzed the archestra enterprise AI platform codebase, working on:
- Task Queue Systems — Poll-based workers with row-level locking, exponential backoff
- Knowledge Base Pipelines — Vector search, embedding, reranking, hybrid retrieval
- MCP Server Runtime — Kubernetes deployment management for MCP servers
- Cache Architecture — Distributed (PostgreSQL) + in-memory LRU caching
- Agent Tool Assignment — Dynamic MCP tool allocation with RBAC
I write clean, maintainable code and understand how to build systems that scale.
| Language | Proficiency |
|---|---|
| Python | Advanced — AI pipelines, Whisper, translation, TTS |
| TypeScript | Advanced — Node.js backend, type-safe APIs |
| SQL | Proficient — PostgreSQL, complex queries, optimization |
- LLM integration (OpenAI, Anthropic, local models)
- Whisper speech-to-text (local inference)
- Edge TTS voice synthesis
- Vector databases & embeddings (pgvector, chunking strategies)
- RAG pipelines, hybrid search (BM25 + vector)
- Prompt engineering & optimization
- Node.js / Express / Fastify
- PostgreSQL + Drizzle ORM
- Redis / Keyv caching
- Kubernetes + Docker
- GitHub Actions CI/CD
- FastAPI, Next.js
- MCP (Model Context Protocol) servers
- Agent frameworks (LangChain-compatible)
- API gateway design
archestra (Fork)
Enterprise AI Platform with guardrails, MCP registry, gateway & orchestrator
Analyzed and contributed to production codebase including:
- Task Queue (
task-queue.ts) — 877 lines, 5 handlers, poll-based worker - Knowledge Base — Chunking, embedding, hybrid search, reranking
- K8s Runtime — MCP server deployment manager
- Cache Manager — Distributed + LRU caching layers
Skills: TypeScript, PostgreSQL, Kubernetes, AI/LLM
Automated YouTube Shorts creation pipeline
End-to-end video processing system:
- Whisper local transcription (free, no API needed)
- Google Translate → Indonesian subtitles
- FFmpeg subtitle burning
- YouTube duplicate detection
Skills: Python, Whisper, FFmpeg, YouTube API
AI-powered video dubbing pipeline
Multi-language dubbing system:
- Chinese video → Indonesian voiceover
- edge-tts neural voice synthesis
- Audio mixing with background music
- SRT subtitle generation
Skills: Python, edge-tts, FFmpeg, audio processing
| Project | Contribution | Status |
|---|---|---|
| archestra-ai/archestra | Task Queue bug fixes | Pending review |
| archestra-ai/archestra | Cost optimization fixes | Pending review |
| archestra-ai/archestra | KB search feedback system | Closed |
(Coming soon)
- Email: 64lamei@gmail.com
- GitHub: github.com/64johnlee
- LinkedIn: (Add your LinkedIn)
- Timezone: Singapore (GMT+8)
⭐ Open to interesting projects and collaboration ⭐