AI Research & Product Engineer | Machine Learning | Deep Learning | LLMs | Computer Vision | AI Agents | Embedded AI

I am still alive. I'm just an ordinary person living and working in an ordinary world, but I am deeply passionate about computers, space, science, engines, robots, artificial intelligence, books, and the future of human-machine collaboration.

I want to help shape a world, an era, and an environment where humans and artificial intelligence can coexist under the same sky. My dream is to support blind people, people with disabilities, people on the autism spectrum, and anyone facing barriers to technology by using AI to make their lives better. (❁´ ◡` ❁)

I research and build AI systems and products, with a focus on training, fine-tuning, adapting, and improving AI models. My work spans agentic AI, OCR, computer vision, speech & audio AI, workflow automation, Edge AI, intelligent systems, and turning research into practical products that create real-world value.

I enjoy working across the entire AI lifecycle — from exploring ideas and developing models to optimizing, deploying, and integrating them into useful real-world systems.

Currently, we are developing and experimenting with pioneering foundation and frontier AI models across language, speech, and vision, with the ambition of making our mark on the global AI map.

Portfolio | LinkedIn | X | Email

• Developing and experimenting with foundation and frontier AI models across language, speech, and vision.
• Training, fine-tuning, adapting, evaluating, and improving AI models for real-world applications.
• Building agentic AI systems with reasoning, routing, tool use, memory, approval, and human-in-the-loop control.
• Researching and applying LLMs, speech & audio AI, computer vision, OCR, multimodal AI, and Edge AI.
• Optimizing models and intelligent systems for edge devices, embedded hardware, and production environments.
• Turning AI research into practical models, systems, products, open-source tools, and reusable research resources.

Personalized edge AI translation runtime for multilingual operations, designed around local-first inference, terminology control, safety checks, and real-time workflow support.

Agentic document intelligence platform for OCR, invoice/document review, risk detection, provider routing, repair loops, and structured document workflow automation.

ROS catkin autonomous driving stack for Bosch Future Mobility Challenge 2024/2025, covering lane detection, ADAS simulation, path planning, embedded optimization, and vehicle control.

AI control layer for company operations, routing high-risk operational requests through parsers, rules, small models, strong models, and human approval to reduce cost while preserving governance.

AI companion research project focused on safe child-facing interaction, memory design, supervision, education support, and responsible AI product experience.

More work: michaeltechs.xyz/work

AI / Data

Python, OpenCV, OCR, RAG, LLM workflows, agentic systems, model routing, computer vision, data preprocessing, evaluation, safety checks, and human-in-the-loop review.

Software

TypeScript, JavaScript, React, Next.js, Node.js, REST API, PostgreSQL, MongoDB, SQLite, Docker, VPS deployment, GitHub workflows, and full-stack product architecture.

Edge / Robotics

ROS, catkin workspace, Raspberry Pi, Jetson Nano, embedded optimization, camera pipelines, serial communication, sensor integration, control logic, and real-time constraints.

• Website: michaeltechs.xyz
• Email: navinservicesnv000@gmail.com
• LinkedIn: minh-cuong-bui
• X: @mih_cuog