Thang Le Viet
I build coding agents and codebase-intelligence systems. Five years shipping production ML and LLM products, most recently at FPT AI Center.
- Most recent role
- AI Engineer at FPT AI Center, AI4SE Lab
- Focus
- Coding agents, codebase intelligence, and LLM evaluation
- Location
- Ho Chi Minh City (UTC+7)
About
My work is about making AI useful inside a real software engineering workflow.
That means understanding a codebase, finding the right context, making a change, and knowing whether it is actually correct.
Selected work
#2 by commits opendev A terminal-native coding agent for parallel, multi-model development workflows. Rust, 751 stars, 86 forks
Experience
AI Engineer
FPT AI Center, AI4SE Lab
Led a LangGraph-based multi-agent system for code review and bug localization across 50+ repositories. Built LLM evaluation pipelines, MCP servers, and production Python backends.
AI Engineer
Resonance Technology
Shipped LLM features that increased platform engagement by 83.7%. Made SERP clustering 40% faster with more than 5× lower memory usage.
Data Scientist
Be Group
Improved top-recommendation CTR from 55% to 70% for Be Food and increased recommendation clicks by 8% through personalization.
Research Assistant
NLP@UIT Lab
Built COVIDROP, the first Vietnamese numerical-reasoning reading-comprehension dataset, and implemented NAQANet and NumNet baselines.
Notes
Research & education
Numerical reasoning reading comprehension on Vietnamese COVID-19 news: task, corpus, and challenges.
Read publication ↗M.Sc. Computer ScienceUIT, VNU-HCM, 2023–2026. GPA 3.7/4.0
B.Sc. Data ScienceUIT, VNU-HCM, 2018–2022. Top 1 in Major
Working knowledge
AI systemsLLM agents, LangGraph, RAG, MCP, tool calling, evaluation, LoRA
EngineeringPython, TypeScript, Rust, PyTorch, Transformers, FastAPI, Docker, Kubernetes
LanguagesVietnamese (native), English (C1)
Let’s work together.
I’m open to AI engineering roles, open-source collaboration, and difficult problems in software development.