After hours I'm having fun creating an AI Platform using my skills in full-stack development and to leverage the interest I've had in AI since my CS masters program.
Google / NVIDIA / AWS / MS / others have amazing systems and capabilities. Most of us have our own GPUs and integrated NPUs at our disposal. I want easy and local access to their capabilities.
I'm a minnow surrounded by 1000lb gorillas, doing my best, and being terrible at analogies.
My general goals are:
Fully on-prem
Federated with other deployments
Trust
Sharing of training data and models
Resource discovery
Annotated inference endpoints for programmatic consumption
I see value in democratized access to AI models and tools!