HACKATHON - Agent Eval Harness

At a glance

An in-person London hackathon focused on an agent evaluation harness. It is a practical option for teams and builders who want to spend an evening working on how AI agents are assessed, rather than attending a talk-led event.

HIKOMORE’s take

The title points to a narrow, technical problem: evaluating agent behaviour. That makes this more relevant to founders and product teams already building with agents than to people looking for a broad introduction to AI. With no published speaker, ticket or agenda detail in the verified snapshot, confirm the expected level of participation before committing.

Best for

• Technical founders developing AI-agent products
• Engineers and product builders working on agent evaluation or testing
• Startup teams assessing how to measure reliability in agent-based workflows

Why consider it

• The format is confirmed as a hackathon, suggesting hands-on participation rather than a conventional presentation programme.
• Its focus on an agent eval harness is relevant to teams treating evaluation as part of AI product development.
• It takes place in person in London, which may suit locally based teams seeking a focused evening session.

Things to check

• Whether attendance requires registration and whether there is a charge, as ticket status and pricing are not confirmed.
• What participants should bring, including whether a laptop, existing project or technical setup is expected.
• Whether the session is suitable for your current technical level and familiarity with AI agents.
• How the hackathon is structured, including any planned outputs, team format and evaluation criteria.
• Venue access and accessibility arrangements, which are not listed in the verified information.

Original FAQ