Back to services
AI Development
Applied AI systems built on real infrastructure — local and hosted LLM inference, retrieval, and structured knowledge, not prompt-only prototypes.
Start a projectOverview
I build applied AI systems that go past a single prompt: retrieval, memory, and inference infrastructure designed to hold up under real usage, not just a demo.
How I Build This
- Local LLM inference (e.g. Ollama-served open-weight models like Qwen2.5) where cost, latency or data control rule out a hosted-only API.
- Retrieval-augmented pipelines combining vector similarity search with structured, graph-based knowledge where relationships matter, not just similarity.
- Clear separation between the inference layer, the knowledge/memory layer, and the application layer, so each can evolve independently.
- Production concerns from day one: containerized deployment, observability, and evaluation before a system runs unsupervised.
Who This Is For
Teams that need AI capability embedded in a real product or workflow — not a proof-of-concept that never survives contact with production data.
FAQ
Do you only work with hosted APIs like OpenAI/Anthropic?
No — I also build with locally-hosted open-weight models via Ollama when cost, latency or data-residency requirements call for it.
What makes this different from a prompt-engineering consultancy?
The focus is systems architecture — memory, retrieval, orchestration — not just prompt wording.