
AI Engineering Lab
From prompt to GPU. No vaporware.

AI is in house, not in the cloud.
Custom SFF workstation. MiniPC for self-hosted services. Dedicated AI server with optimized thermals.
Bare-metal inference. No API dependency.
docker • n8n • postgresql • redis • minio • ollama • vllm • comfyui

Autonomous rover with vision and voice.
Perception → Decision → Expression → Motion → Loop.
Test scenario: locating my cat.
m5stack • esp32 • c++ • n8n • ollama • vision models • tool-use • tts

Personal knowledge assimilation system.
Multi-source ingestion, relevance-scored takeaways, weekly digests, and semantic search agent.
n8n • postgresql • pgvector • rag • openai • tool-use

comfyui • gradio • n8n • diffusion models • vision models

langgraph • context engineering • tool-use • postgresql • redis

Real-world execution experiment in AI-assisted coding.
From zero to production in ~30 hours.
python • django • postgresql • tailwind • heroku • openai • anthropic