Data Decoded LDN ran over two days at Olympia London, and I came away with more to think about than I expected. It did not start well. A London train strike turned my commute into a nightmare and I missed the early keynote on day one entirely, but the sessions I did catch made the effort worth it.

Day one

Once I got there, day one was packed. I sat in on sessions across data engineering, AI automation and analytics, and there was plenty to chew on in each of them.

See the day one post on LinkedIn

Day two

I spent a good part of day two on the exhibition floor, at the Oracle, Experian, Hex, Amazon Web Services and Fivetran stands. Good conversations across the board.

One that stood out was with Darren Lynch, an Oracle representative, about their latest products and how smaller companies can use Oracle infrastructure to scale. Seeing Oracle still operating as a powerhouse took me back to my university days, when their monthly magazine would arrive and I would actually look forward to reading it.

See the Oracle conversation post on LinkedIn

The two talks that stayed with me

Matheus Guimaraes, Senior Developer Advocate for the UK and Ireland at AWS, talked about building agentic architecture on Amazon Bedrock, the managed AWS service that gives secure, API-driven access to foundation models from the leading AI companies. The part that stuck with me was his honest acknowledgement that AI success depends on your data foundation, not on which model you pick.

Ruth Buscombe, the Formula One race strategist, was the standout. Her point was simple but sharp: every team on the grid has access to the same data. What separates winning from losing is what you do with it and how quickly you communicate the decision. That framing transfers straight into data work, and I left thinking about it more than anything else from the two days.

See the wrap-up post on LinkedIn

Thanks to the Data Decoded LDN team for a well-organised two days. I am already thinking about what to take back into my own work.