Building AI Products on the Right Foundations
For the last couple of years most of the AI conversation has been about models: which one to use, how to prompt it, how to wrap it in an agent. As more of those products reach real users, the conversation inside teams is shifting to what the model is actually working with.
An AI product is only as good as the data feeding into it. A support agent is only as useful as the tickets and call transcripts behind it. A contract review tool is only as accurate as what it pulls out of the PDFs it reads. Getting that right has become one of the hardest and most important parts of building with AI.
That puts Data Engineers and AI Engineers closer together than they have ever been. Data teams are working with far more unstructured data than before, documents, emails and transcripts, and are often responsible for the pipelines that feed AI products directly. AI Engineers are building the models and agents that rely on that data, and working out what to do when it is incomplete, inconsistent or not there at all.
Both are working on the same problem from different ends. This session brings them into one room to talk about what good data for AI looks like, and how teams are building it.
Aura Talks are small, focused sessions bringing together practitioners working through these problems first hand.
📋 The Session
We’ll hear from experts working across the data and AI sides of the stack, each giving a short talk from their own experience before opening it up to the room.
Kai Loh, AI Manager at 9fin, the AI-native intelligence platform for capital markets. More details on Kai’s talk to follow.
Speaker 2 - TBC
Topics we’ll explore include:
What makes data ready for AI, and how that differs from data built for analytics
Working with unstructured data at scale: documents, emails and call transcripts
How you measure quality when there is no schema to test against
What AI Engineers need from the data behind their products, and what they do when they can’t get it
Tracing a model’s output back to the data it came from
How data and AI teams work together, and where ownership sits
The session will be discussion-led rather than presentation-heavy. Each speaker will share their experience before opening up the conversation to questions and discussion from the room.
⏰ The Format
6:00pm - Pizza & drinks
6:30pm (sharp) - Lightning talks & Q&A
7:30pm onwards - Networking
👥 Who’s in the room
AI Engineers and ML Engineers building AI products and agents, alongside Senior, Staff and Lead Data Engineers and Data Platform Engineers, from some of London’s most exciting startups and scale-ups.
This is a small, invite-only evening designed for the engineers building the data and the products that depend on it.
Places are limited to keep the discussion interactive and high quality.