01 · AI PlatformIn progress
AI Release Quality Gate
An automatic LLM evaluation that runs on every prompt or model change and blocks the release if quality drops.
LLM evaluationCI/CD for AIQuality standards
AI Platform Product Manager · Madrid · Open to EU, UK, Canada, US
I turn enterprise data platforms into AI products that ship.
9+ years building enterprise data and cloud platforms at IBM, Infosys and Genpact. Most recently I led a GenAI product on Amazon Bedrock from requirements to production and owned 70+ data and AI initiatives. MBA, IE Business School.
3x
faster product delivery with a GenAI product I led on Amazon Bedrock
70+
enterprise data and AI initiatives led across 10 cross-functional teams
$600K
annual revenue from the AI and data portfolio I owned
AI Product Lab: one online shop, four AI decisions
Four hands-on builds for the same European online shop. Two for the AI platform other teams build on, two for AI features customers use. Each one answers: is it good enough to launch, at what cost, and with what risk?
02 · Customer AIUp next
Support Agent with Human Approval
An AI agent that resolves order and refund requests, asks a human before risky actions, and is protected against data leaks and prompt attacks.
AI agentsGuardrailsA/B testing
03 · AI PlatformUp next
Model Routing & AI Cost Optimizer
Routing each request to the cheapest model that is good enough, with a spend-by-team dashboard and an annual savings case.
Model routingInference costBuild vs buy
04 · Customer AIUp next
AI Shopping Assistant
Answers shopper questions from the product catalogue and reviews with sources, says “I don't know” when unsure, and is tested for conversion.
RAGHallucination testingAI UX
How I work
Start with the workflow, not the model
I map where people lose time before choosing any AI.
Trust is a feature
Every AI output is evaluated before anyone relies on it.
Measure adoption, not demos
Success is teams and customers using it every week.