About / The engineering notebook
About Viet-Anh
ML Lead focused on edge AI, computer vision, and privacy-first LLM systems. 5.6k+ stars across open-source work.
01 / Profile
What I work on.
Machine Learning Lead at Scopic Software and Founder and maintainer at Neural Research Lab. I build production ML systems across both Cloud and Physical AI — cloud-served LLMs to on-device inference on edge hardware. Focus areas: local and privacy-first AI, high-performance computer vision, and AI security.
End-to-end across the stack — dataset design, model training, optimization, deployment, monitoring. The goal is tools developers actually reach for, running on real hardware at real cost.
Currently: leading ML at Scopic, building AnyLabeling and local-first AI systems, and publishing measured engineering studies at vietanh.dev. I am also pursuing an M.S. in Computer Science at The University of Texas at Austin (expected 2029).
Open Source
Projects I lead and maintain, accumulating 5.6k+ GitHub stars:
- AnyLabeling: AI-assisted image and video labeling with local Segment Anything and YOLO inference. 3.5k+ stars.
- EdgeVox — On-device voice-agent framework. ROS2-native, multi-language, safety-first agent architecture. Targets robots, desktop apps, and embedded systems.
- Bulwark: Local Linux security scanner with reversible hardening, file-integrity monitoring, and checks for secrets exposed to coding agents.
- Llama Assistant: Maintained privacy-first offline AI desktop assistant with local models, RAG, voice, and multimodal input.
- Nôm VN: Vietnamese NLP package with license-tracked evaluation data, retrieval, OCR, and diacritic-aware metrics.
- DaisyKit — Cross-platform AI SDK for computer vision and ML. 300k+ PyPI downloads.
Technical writing on vietanh.dev covering edge AI, agent reliability, AI security, and the engineering trade-offs behind production systems.
Focus Areas
- Local-first, privacy-respecting AI applications
- High-performance computer vision and perception
- Edge deployment (TensorRT, NCNN, Jetson, llama.cpp)
- AI security and production MLOps
Credentials
- AWS Certified Machine Learning — Specialty (2025)
- NVIDIA-Certified Associate: Generative AI LLMs (2024)
- IEEE BigData 2023 — Active learning for object detection
02 / Skills
Focus and toolkit.
Focus
- Edge AI
- On-Device Agents
- Computer Vision
- Large Language Models
- Privacy-first AI
- MLOps
- AI Security
Stack
- Python
- C/C++
- JavaScript
- TypeScript
- PyTorch
- TensorRT
- NCNN
- llama.cpp
- ONNX
- Docker
03 / Elsewhere
More of the work.
04 / Journey
Career timeline.
April 2024 - Present
Scopic Software
After ending my journey at Samsung SDS, I joined Scopic Software as a Remote Machine Learning Engineer, which has been a great opportunity to work with a global team and solve new challenging problems. Now leading the ML/AI team as the first Machine Learning Lead.- Jul 2025 - Present: Machine Learning Lead - The first Machine Learning Lead at Scopic. Overseeing ML/AI Projects and collaborating with cross-continental teams to build innovative AI-powered software solutions that meet the diverse needs of our global customer base. Conducting extensive research on cutting-edge machine learning techniques and algorithms. Performing technical interviews for Machine Learning Engineer candidates.
- April 2024 - Jun 2025: Senior Machine Learning Engineer.
Jun 2026 - Expected Mar 2029
The University of Texas at Austin
Master of Science in Computer Science
Pursuing graduate study alongside my work as Machine Learning Lead. I am using the program to deepen the theoretical and research foundations behind the production AI systems I build. Program details.
Oct 2021 - April 2024
Samsung SDS R&D Center Vietnam
- Architected the system and developed key backend modules in the AI Core MLOps project – a no-code platform for labeling, training, and deploying AI solutions for vision inspection, OCR, and NLP (14 AI use cases).
- Trained models and developed algorithms for inspection applications, deployed successfully in Samsung's factory.
- Awarded: AI Research Lab "The best employee of the year" 2022.
- Awarded: Samsung SDS Hanoi "Employee of the month - April 2022" - recognized as "Key person to design MLOps architecture of Al-based Inspection System".
Oct 2020 - Oct 2021
VinGroup (VinAI + VinBigData)
- Joined Autopilot project at VinAI Research as an AI Engineer. Automated the data collection process, contributed to hardware, localization & mapping, LiDAR research, and a self driving car demo of Vingroup.
- Oct 2020 - Feb 2021: Participated in Vingroup AI Training Program for 120 engineers. Leader of a 4-engineer team to build 2 open source projects: Push-up Counter and VN AI Doctor.
Sept 2019 - Oct 2020
Kaopiz Software AI Team
- Built AI core for Image Processing, OCR, object detection, and segmentation.
- Optimized and deployed models for servers, desktop apps, mobile apps, and embedded systems.
2015 - Aug 2020
Education
Went to Hanoi University of Science and Technology, one of the best universities in Vietnam to study Computer Science. Here I established this blog at vietanhdev.com (changed to vietanh.dev later).
- The best presentation award for thesis, later my project was recognized by NVIDIA as The Jetson Project of the Month October 2020.
- Runner Up - SoICT - IBM Hackathon 2020.
- Teaching assistant certification for a database course
- The First Prize - FPT's Driverless Car Challenge - University round.
- Global Project-based Learning Scholarship - Short-term robotics course in Tokyo - Japan.
2009 - 2014
Self-taught programming
Learnt programming at 12 on an Intel Pentium 4 with 1GB RAM. First program was in AutoIT — it calculated the perimeter of a triangle from three measurements. Wrote a hard-drive manager the same year. Then dropped it.
Came back in 2014 with a chess program that used the Spike chess engine to play against humans — also in AutoIT. Between those, I just dreamed of becoming a computer engineer.