Consulting
AI Product Development
I build LLM features that survive contact with real users — not just demos. From scoping the right feature to shipping it with evals, guardrails and a cost budget, I've done this on my own products and can do it on yours.
LLMsRAGEvalsGuardrailsPrompt engineering
Discuss this
What you get
- Feature scoping that separates what an LLM should do from what plain code should.
- Fast prototypes so you can feel the feature before committing to a full build.
- RAG pipelines that ground answers in your own content when accuracy matters.
- Eval suites and guardrails so quality is measured, not hoped for.
- Cost and latency budgets, plus monitoring once it's live.
How it works
- Scope Define the feature, the success criteria, and where the model must not be trusted.
- Prototype Build a working slice quickly to validate the approach with real inputs.
- Harden Add RAG, evals, guardrails and cost/latency budgets before launch.
- Ship & monitor Release, watch quality and cost in production, and iterate.
A good fit if
- You have an AI feature idea and want it to reach production, not stall as a demo.
- Accuracy, cost or safety matter enough to need evals and guardrails.
- You want someone who's shipped LLM products, not just prototyped them.
FAQ
Have you actually shipped AI products?
Yes. Luna generates personalized children's stories and Smart Clip rewrites text across platforms — both live, both built by me end to end.
Can you work with our existing codebase?
Usually, yes. I'll audit what's there first so the AI feature fits your stack instead of fighting it.
How do we start?
Through the form and a short intro call to scope the feature. Engagement shape and timeline depend on what we're building.
Contact
Discuss AI product development
Tell me what you're building. Remote-first from Barcelona — we start with a short intro call.