I build practical systems around computer vision, video analytics, and edge AI, with a strong focus on turning algorithms into reliable, deployable software.

My background is primarily in C++ video and vision systems, including real-time video pipelines, object detection and tracking, and edge inference.

I'm currently exploring and building projects around:

• Video Intelligence — detection, tracking, video understanding, and privacy-aware video processing
• Edge AI — efficient inference on embedded and edge platforms
• Vision + VLM — bringing modern vision-language models into practical video applications
• AI-assisted systems — combining traditional computer vision with LLM/VLM-based reasoning
• IoT & local intelligence — lightweight sensing and automation systems that work locally

Core

C++ · Python · OpenCV · GStreamer · DeepStream · Docker · Linux

AI / Vision

Object Detection · Multi-Object Tracking · Video Analytics · Edge Inference · VLM

Currently exploring

Local LLMs · AI Agents · Embedded AI · Nvidia Jetson · RK3588 · Home Automation / IoT

I’m particularly interested in projects that sit between research prototypes and real-world systems — small enough to understand, deploy, and maintain, but useful enough to solve an actual problem.

Some recurring themes in my work:

• real-time video processing
• privacy and security
• efficient edge deployment
• local-first AI
• practical open-source tooling

I'm building open-source projects through @Nanexus-AI, with an initial focus on practical tools for computer vision, video intelligence, and edge AI.

I'm interested in contributing practical tools to the open-source community, while also exploring how open technologies can evolve into reliable solutions for real-world and commercial applications.

I'm open to collaborations involving custom development, system integration, and applied computer vision / edge AI projects.

If you're working on an interesting real-world problem in these areas, feel free to get in touch.