Data products · analytics · data and AI governance

<aside> 💡

→ Open the interactive portfolio

</aside>

Can you trust this data enough to decide with it, or to let an AI agent 🤖 act on it?

I've spent 10+ years in analytics making numbers trustworthy. Now I build data products on top of them: AI agents, LLM features with evals, and a person in the loop where it matters. Every number below comes with how it was measured and what it can't tell you.

• Looking for: senior and lead roles in analytics, data products, and data and AI governance
• Experience: 10+ years in analytics across the UK and Spain: retail, travel, UK government, industrial and health tech
• Based in: Europe, working remotely. EU passport and UK Settled Status
• Contact: CV · LinkedIn · Blog

From trusted data to useful AI

Analytics taught me when a number can be trusted. An AI agent needs the same answer before it acts on data, so that's where I start.

• Build · Data products. Co-founder and AI product manager of Nutr(IA) Coach, from concept to beta. I designed the agent architecture and data flows with LLM APIs and Vertex AI on Google Cloud, and I run evals on the LLM features.
• Govern · Who checks the AI. We gave the AI to nutritionists, not the public, so a professional reviews every recommendation before it reaches a client. I'm now studying AI Cybersecurity and Governance with Columbia Engineering.
• Use · AI in the analysis. At JD Sports I used the OpenAI API to categorise open-text survey answers at scale. Day to day I use AI (Claude Code, local models with Ollama) to automate repetitive tasks, so my own time goes on verifying the output before I sign it off.

How I design it: Client data → Can it be trusted? → 🤖 AI agent drafts → Nutritionist reviews → Client

Case 01 · A checkout that looked broken (+4% checkout conversion)

JD Sports Group (Sprinter and Sportzone) · Senior Digital & Marketing Data Analyst · 2021–2024

The business wanted the checkout redesigned. Before anyone committed engineering time, I compared the funnel step by step with the retail benchmark, and found a page carrying 40% of checkout traffic that had never been tracked. I fixed that first.

Three teams wanted three different fixes. Marketing wanted Customer Care on video calls, Customer Care wanted chat on product pages, and the evidence pointed at the checkout. I showed each team what data existed at each stage and what each option would tell us, and we started with the checkout.