About

What I do now

I am currently Tech Lead for the UK Government, leading the technical direction of a strategic data-integration programme. Most days the work is less about models and more about systems. I connect modern cloud data platforms to legacy ones that were never meant to talk to each other. I design a delivery model that takes an idea through to production support, and I set the engineering standards and AI governance that let a regulated organisation adopt new tools without losing its footing. I have also been bringing AI-assisted engineering into the team's day-to-day work, and prototyping agent-based tools with evaluation built in from the start.

Outside the day job I build products of my own. Benchd is a voice mock-interview coach in private beta. Poppy is a multi-agent receptionist for UK private clinics, still a working demo. AutoMeal, which I co-founded with José Enrique, plans meals from what is already in the pantry. FastAPI Rocket Boilerplate is an open-source starter that has picked up a few hundred stars and a life of its own. The longer write-ups are under Projects.

How I got here

I studied Telecommunication Systems Engineering and Business Administration at the same time at Universidad Rey Juan Carlos in Madrid. That double track never really left me. I care about how a system works, and just as much about whether it is worth building. A postgraduate programme in technological change and innovation at BBVA followed, and so did my first job, building automated reporting and alerting for financial derivatives.

The next few years were in banking and fintech: a KYC and anti-money-laundering tool at Santander, and lending platforms at Aplazame. That is where I learnt that in a regulated environment, trust comes from data quality and traceability, not from clever code. Then came startups and scale-ups. I built the game backend of a connected toothbrush at Playbrush. At Senniors I hired and led a team of eight to build an active-ageing app with partners including Google, Fitbit and Zurich. At Erlang Solutions I worked on invoice-reading OCR and computer vision, at Climate Policy Radar on language models for climate-policy analysis, and at Monolith AI on fine-tuned models.

More recently I was a technical lead at British Airways, leading seven engineers and six data scientists on generative AI for flight planning, and briefing the COO's team on the work. After that I was an AI engineer at Coinbase, designing multi-agent systems and production language-model services, where security, evaluation and failure recovery were part of the job. I have moved back and forth between hands-on engineering and leadership on purpose. Each one makes me better at the other.

In February 2026 I joined a panel at the AI & Big Data Expo Global in London, on scaling AI across the enterprise: how to get from a pilot to real adoption without taking on unnecessary risk. That question sits under most of what I write here.

What I write about

My working belief is that the bottleneck is rarely the model. AI projects stall when the purpose gets lost, when ownership is vague, or when the people building the system sit too far from the people who have to use it. So the essays are about three things.

• Production AI. What it takes for an agent or a retrieval system to survive real users, real cost and real regulators.
• Leading engineering teams. Mentoring, engineers and data scientists in the same room, and keeping a team pointed one way when the work is still ambiguous.
• Technology as a business decision. What to build, what not to build, and how to explain the trade-off to someone who does not write code.

Outside the job

I read widely outside the work: philosophy, history, economics, neuroscience and politics, because incentives decide whether a system is still running a year from now. I am learning Chinese, Russian and Arabic, since the language you have changes the problems you can see. London is home. I still spend time in other technology hubs, including Dubai, the US and Singapore.

If something here is useful, or you want to argue with it, the page is Contact. A CV is here if you want the longer record.

Education

• Telecommunication Systems Engineering, Universidad Rey Juan Carlos, 2013–2018. 2:1.
• Business Administration and Management, Universidad Rey Juan Carlos, 2014–2018. 2:1.
• Technological Change and Innovation, postgraduate, BBVA, 2018. Top of the class, 9.3 out of 10, among 70 students, and captain of a team of 15.