Private & open-weights AI
Most companies don't need to send their data to someone else's API. I help teams deploy local models, wire them into their own documents and systems, and get real work out of them.
- Local and on-prem LLMs: model selection, hardware sizing, deployment
- Open-weights models (Llama, Qwen, Mistral, DeepSeek, gpt-oss) tuned to your use case
- RAG over internal docs, wikis, tickets and databases
- Integrating AI into existing workflows, with evals so you know it works
- Privacy, data residency, and cost vs. hosted APIs