About
AI Engineer at MoMo, based in Ho Chi Minh City.
I specialize in multi-agent systems, LLMs, real-time voice agents and production AI infrastructure. I started building AI in 2021, alongside my Mechatronics Engineering degree at HCMUTE, and have worked as an AI Engineer since 2022. Mechatronics gave me a hardware and embedded-systems side that still shows: Assistant Core runs on ESP32 devices and smartwatches as well as in apps.
Experience
-
MoMo
Dec 2024 – present
AI Engineer
Led end-to-end AI engineering for Trợ Lý Ăn Chơi, MoMo’s AI lifestyle assistant, reaching 1M+ users across 28 in-app entry points. Architected a multi-agent system (1 orchestrator + 6 domain agents) with 90 tools, shipping agentic commerce journeys for cinema, flight, bus and train tickets, hotel and experience bookings, and Highlands Coffee orders through to MoMo checkout. Cut LLM cost to $0.0015 per daily active user with a 95% prompt-cache hit rate, and built a Memory Agent plus an online LLM-as-a-judge pipeline. Won Champion of Hacking Week at MoMo Tech Day 2025.
-
FPT Software – AI Center
Mar 2024 – Dec 2024
AI Engineer
Built an enterprise IT-support assistant that pairs LLM planning with deterministic workflows and long-term memory, and a multimodal RAG pipeline over text, images, charts and tables in PDF and DOCX files with 90%+ accuracy on the project’s evaluation set.
-
FPT Software – Quy Nhon (QAI)
Jun 2022 – Feb 2024
AI Engineer
Built an enterprise document search product for a Japanese client on embeddings and vector search, then extended it into RAG; shipped real-time face recognition to production on AWS. Received the Most Valuable Player 2022 award and won Second Prize at the Quy Nhon AI Hackathon 2022.
-
Hao Phuong Corp
Jan 2022 – Jun 2022
Backend Engineer
Owned the backend of a real-time smart-farm management system on ASP.NET Core: sensor data collection, remote device control and monitoring dashboards.
Education
-
Ho Chi Minh City University of Technology and Education (HCMUTE)
2018 – 2022
Mechatronics Engineering
First Prize for the graduation project defense: an IoT control and monitoring system for an automatic paper-spreading machine. Built my first deep learning projects alongside the degree.
Community
-
Sudo Code – Women Techmakers & GDG Ho Chi Minh City
2025 – present
NLP Mentor
NLP mentor for Sudo Code 2025 and 2026.
Skills & stack
- Languages
- Python, TypeScript, JavaScript, SQL
- LLMs & agents
- OpenAI, Anthropic, Gemini and xAI APIs; LangChain, LangGraph; multi-agent orchestration, tool calling, MCP, context engineering, prompt caching, LLM-as-a-judge evaluation
- RAG & search
- Hybrid search (pgvector + full-text, RRF), embeddings, OCR, Elasticsearch, Qdrant
- Voice AI
- Streaming ASR and TTS, VAD, real-time speech-to-speech, barge-in, Opus
- ML serving
- NVIDIA Triton, TensorRT, gRPC, dynamic batching
- Backend & web
- FastAPI, Next.js, React, Pydantic, SSE, WebSocket, MQTT
- Data
- PostgreSQL, Redis, BigQuery, Airflow
- Observability
- Langfuse, LangSmith, Prometheus, Grafana
- Cloud & DevOps
- AWS (EKS, Lambda, S3, CloudFront), GCP, Docker, Kubernetes, Helm, GitHub Actions, Nginx
What I’m building
Outside of work I build Assistant Core, a platform for deploying one AI assistant configuration everywhere: web, mobile widgets, IoT devices, wearables and embedded surfaces. See what it does.
How I work
I care about shipping real products, not prototypes. My instinct is to go from zero to production as fast as possible, then iterate on real user feedback. I tend to think in systems: how components connect, where failures propagate, what the user actually experiences end to end. I’m most energized when I can own a problem fully, from architecture to deployment.