Post-Training, Evals & Improving Agents - AAIF London x Prolific
Post-Training, Evals & Improving Agents - AAIF London x Prolific
How do we make agents meaningfully better after the base model has been trained?
Join AAIF Community London and Prolific for a technical evening exploring post-training, agent evaluation, feedback loops, reinforcement learning, agent architecture, and the emerging idea of iterative and recursive self-improvement in agentic systems.
AAIF Community London is partnering with Prolific for this event, with Prolific supporting the venue, food and drinks.
As agents move from demos into real-world workflows, improving them becomes a systems problem as much as a modelling problem. Better models matter, but so do the environments they learn from, the evaluations we use, the feedback we collect, the tools and context we give them, and the loops through which agents can learn from previous behaviour.
We’ll bring together researchers, engineers and practitioners working across these problems for four focused technical talks, followed by a speaker panel and audience Q&A.
The first hour will be dedicated to food, drinks and networking before the technical programme begins.
Capacity is limited, so please only register if you plan to attend.
What We'll Cover
We’ll explore topics including:
Post-training techniques for improving model and agent behaviour
Agent evaluation and what we should actually be measuring
Feedback loops, environments and reinforcement learning
Building and improving production agentic systems
Agent harnesses, tools, context and architecture
Learning from traces and real-world interaction
Iterative and recursive self-improvement - and what those ideas look like in practical agent systems
Where current approaches break down and what comes next
Expect technical talks, practitioner perspectives, an open panel discussion and plenty of time to meet other people working on these problems.
Agenda
17:30 - 18:30 | Doors Open, Food, Drinks & Networking
Check in, grab some food and drinks, and meet engineers, researchers, founders and builders working across AI, agents, evaluation and post-training.
18:30 - 18:40 | Welcome & Kickoff
Opening remarks from Agentic AI Foundation (AAIF) Community London and Prolific.
A quick introduction to the evening, the speakers and some of the questions we’ll be exploring around evaluating and improving agentic systems.
18:40 - 18:55 | Technical Talk #1
Talk title TBA
🎤 Speaker TBA
15-minute technical talk followed by a short transition to the next speaker.
18:55 - 19:10 | Technical Talk #2
Talk title TBA
🎤 Speaker TBA
19:10 - 19:25 | Technical Talk #3
Talk title TBA
🎤 Speaker TBA
19:25 - 19:35 | Break
Refuel, grab a drink and continue the conversation before the final talk and panel.
19:35 - 19:50 | Technical Talk #4
Talk title TBA
🎤 Speaker TBA
19:50 - 20:25 | Panel Discussion + Audience Q&A
We’ll bring the speakers together for a hosted discussion on some of the bigger questions raised throughout the evening.
We’ll explore topics such as:
What actually makes an agent improve?
Where do current agent evaluations fall short?
How should feedback from real-world agent behaviour feed back into training?
What belongs in the model versus the agent harness?
How far can iterative or recursive self-improvement realistically go with today’s systems?
What changes as agents become more autonomous and operate across longer time horizons?
What are the most important unsolved problems in post-training and agent evaluation?
We’ll then open the discussion to the audience.
Bring your questions.
20:25 - 21:00 | Networking
Continue the conversation with the speakers, attendees and other people working on post-training, evals, agents and production AI systems.
21:00 - Onwards | Optional Social
For anyone who wants to continue the evening, we’ll head somewhere nearby after the event.
Location: shared on the night.
Who Should Come
This event is aimed primarily at:
AI/ML engineers, research engineers, researchers, agent developers, applied AI teams, infrastructure engineers, founders and technical practitioners working on or interested in:
post-training
model and agent evaluation
reinforcement learning
agentic systems
AI infrastructure
coding agents
feedback and learning loops
production AI systems
You don’t need to be working directly on post-training research to attend. If you’re building agents and thinking seriously about how we evaluate, improve and learn from them, you should find the discussions relevant.
About AAIF
The Agentic AI Foundation (AAIF) is a community dedicated to advancing the understanding and practical application of agentic AI. We bring together engineers, researchers, founders, builders, and AI enthusiasts to explore how autonomous AI systems are designed, evaluated, and deployed.
Through reading groups, workshops, and community discussions, AAIF brings together engineers, researchers, founders, and AI enthusiasts to exchange ideas, challenge assumptions, and learn from one another.
By attending, you agree to our Code of Conduct and Privacy Policy.