Backend & AI Engineer based in Ho Chi Minh City, Vietnam.
I transitioned from logistics operations into software engineering focusing on async Python backends and Applied-AI for production systems.
1 year into my professional engineering journey, including 2 months as a Software Developer in Chandigarh, India.
- Applied Generative AI: Designing production-grade LLM services.
- AI-Native Architecture: Building workflows designed from the ground up around autonomous agents and MCP tool.
- AI System Reliability & Guardrails: Hardening non-deterministic AI behavior using strict validation and AST invariants.
- Deterministic Chaos Testing: Stress-testing agent workflows with simulated latency, fault injections, and replay engines.
1. arc-reliability — Autonomous Reliability & Chaos Harness for AI Agents
Open-source Python library published on PyPI
- An interceptor SDK (
@trace_agent,@trace_tool) designed to trace, test, and inject faults into autonomous agent runs. - Features virtual clock injection, sandboxed fork-replay, and automated delta-debugging (
ddmin) to reduce transient LLM failures into reproducible test cases. - Implements AST invariant mining via Tree-Sitter to verify transactional integrity rules.
- Links: PyPI Package • Source Code
Real-time audio AI service
- Built a bidirectional streaming pipeline decoupling real-time dialogue generation from background evaluation workers.
- Integrated WebSocket streaming with acoustic filler cues to cut Time-to-First-Audio (TTFA) from 3,500ms down to <200ms.
- Links: Live Application Demo
Domain-Driven Warehouse Management Backend
- Designed around real-world warehousing constraints (FIFO rules, slotting optimization, capacity limits).
- Handled high-concurrency checkout and stock movement using PostgreSQL row-level locks and Redis distributed locks to prevent overselling.
- Links: Source Code
| Area | Tools & Technologies |
|---|---|
| Languages & Core | Python (FastAPI, Asyncio, Pydantic), SQL, JavaScript / Node.js |
| Data & Messaging | PostgreSQL, Redis (Streams, Distributed Locks), SQLAlchemy |
| AI & Protocols | Model Context Protocol (MCP), gRPC, OpenTelemetry, AST Parsing (Tree-Sitter) |
| Testing & Tooling | Pytest, Docker, Git/GitHub Actions, Linux/WSL |
Early in my engineering path—focused on turning brittle, non-deterministic AI into resilient production systems. Always seeking tough code reviews, complex failure modes, and opportunities to harden open-source AI tooling alongside experienced engineers.