Projects
Work worth pointing at.
A small portfolio of the things currently carrying the most signal: industrial AI, local inference, hardware, and applied reliability.
-
shipping
Livit AI
Enterprise AI consulting and product company for industrial reliability, experiential AI systems, and GPU expansion hardware.
Industrial operators do not need another generic chatbot. They need AI systems that understand uptime, safety, privacy, and the practical limits of deploying useful tooling close to the work.
commercialAI infrastructureindustrial -
in development
GPU Bridge
Patent-pending external multi-GPU expansion hardware for workstations and on-prem inference.
The bottleneck for useful local AI is often physical: power, slots, thermals, cabling, and the cost of scaling compute without rebuilding the entire workstation around it.
hardwareAI infrastructurePrivate for now -
shipping
Local LLM project
Personal exploration of on-device LLM deployment and the tooling around it.
A lot of AI value will live outside hosted APIs. Local inference forces sharper thinking about model size, context, latency, privacy, and where the system actually runs.
personalLLMtooling -
in development
Tinkering with robots
Personal robotics experiments using 3D printing, 3D modeling, a Comma 4, and a Comma 3.
Robotics is where software has to negotiate with geometry, sensors, power, fabrication limits, and the physical world. The interesting part is not a polished demo. It is the loop between CAD, printed parts, perception hardware, control, and everything that breaks in between.
personalroboticshardware3D printing