​How do you know when an AI agent is reliable enough for production, and where do you look when it fails?

​Join AAIF Community London, Braintrust and Modal for a hands-on technical workshop on evaluating agent behaviour, investigating traces and improving reliability.

​Bring your laptop.

​What you'll work through

• ​Build evals that identify agent failures

• ​Investigate agent behaviour through traces

• ​Run eval workloads in Modal serverless sandboxes

• ​Measure and improve agent quality with Braintrust

• ​Explore how evals fit into production agent development and reliability workflows

​The emphasis is on hands-on implementation, debugging and evaluation practice.

​Agenda

​18:00 - Doors, registration & networking
18:30 - Welcome from AAIF London
18:40 - Technical introduction from Braintrust & Modal
19:00 - Hands-on workshop
20:30 - Drinks, canapés & networking
22:00 - Close

​Who should come?

​For senior engineers and technical practitioners working with AI agents, evals and production AI, including:

​AI/ML engineers · Applied AI engineers · Research engineers · Agent developers · ML/platform engineers · Technical founders

​AAIF London events are curated and registration approval is required.

​About AAIF, Braintrust & Modal

​Braintrust is an evaluation and observability platform for AI applications and agents.

​Modal provides serverless infrastructure for AI and ML workloads.

​AAIF Community London is part of the Agentic AI Foundation under the Linux Foundation, bringing together engineers and researchers working across agent infrastructure, interoperability, evaluation, developer tooling and production AI.

​AAIF community events are technical, practitioner-led and vendor-neutral.

​Registration & privacy

​Capacity is limited.

​Attendee information remains on the AAIF side unless you explicitly opt in to hear from an event partner.

​By attending, you agree to the Linux Foundation Code of Conduct and Privacy Policy.