The F1 Playbook, Chapter 2:
How to Build a Test Data Search Engine

When something goes wrong in Formula 1, engineers do not have the time to go searching through thousands of files and channels. Decisions have to happen at race pace.

AI can cut through the noise to identify patterns and anomalies at speed. But simply asking AI to compare a suspect run against every similar run from the last six months will fail. Recordings aren’t linked to the build, the software version, or the conditions they came from.

So how are F1 organisations structuring data so that an engineer can ask a question in plain language and get an answer back?

On 29th September, Michael Rosam and Tomáš Neubauer will demonstrate how to structure engineering data with the right context, metadata and relationships, so that AI agents can investigate engineering problems more effectively.

Sign up and find out how to:

• Get thousands of channels from multiple vehicles into one searchable place
• Keep the build, software version and conditions attached to every recording
• Compare a new run against months of history without opening files one at a time
• Let AI agents write the analysis to fit the question being asked
• Judge whether an AI answer can be trusted before acting on it