Upcoming · December 10, 2026 · Sunnyvale, CA

Agentic AI × RL

Research Talk and Demo Night

1 research talk · 8 talks & demos · A full hour of networking

Agentic AI × RL is an evening for the researchers, founders, engineers and investors working on agents that get better the more people use them. One research talk, eight live talks and demos, and a curated room of about 80 people in Sunnyvale.

Apply on Luma A curated room of 80 — every application is reviewed.
Address shared with approved guests
Agentic AI × RL: Research Talk and Demo Night — event cover
01

Why now

Reinforcement learning taught models to reason. The next problem is teaching agents to learn from the people they work for. Every correction, approval, edit and return visit is a learning signal, but it is implicit, sparse, noisy and personal, and turning it into a better agent without forgetting or regressing is still an open problem.

02

What we are exploring

  1. 01Reward from people: corrections, approvals, edits and retention as learning signals
  2. 02Online and continual learning, without forgetting or regressing
  3. 03Personalization: one policy for everyone vs. adapting to each user, and cold start
  4. 04User simulators and environments for training and evaluating agents
  5. 05Reward hacking and sycophancy: serving the user's goals, not their clicks
03

Agenda

  1. 16:30–17:00Doors open, check-in and networking
  2. 17:00–17:20Opening
  3. 17:20–18:05Round I · Agents in the Wild (4 teams)
  4. 18:05–18:15Break
  5. 18:15–19:00Round II · The Learning Stack (4 teams)
  6. 19:00–20:00Networking
04

Who should join

  • RL researchers and engineers working on RLHF, RL from implicit feedback, online and continual learning, personalization, reward modeling or post-training
  • Founders and product teams with agents in production: personal agents and assistants, browser and computer-use agents, and agents that run real workflows
  • Teams building the training stack: user simulators, environments, evals, and the inference and compute behind per-user learning
  • Investors backing RL, post-training and consumer AI
05

FAQ

Who should apply to attend?
RL researchers and engineers, founders and product teams with agents in production, teams building user simulators, environments, evals and the compute behind per-user learning, and investors backing RL, post-training and consumer AI. It is a curated room of 80 and every application is reviewed — tell us what you're building or researching so we can make sure the right people meet.
How do I apply to speak or demo?
Email partnership@openstagesai.com. Tell us whether you want to give a talk or a demo, and what your agent or research learns from. Eight teams present, four per round, with 8 minutes each plus 3 minutes of questions. We're looking for teams whose agent gets better from how people use it, or who build the RL that makes that possible. A polished deck isn't needed — bring a demo or something to show the room.
Can my company partner with the event?
Yes. Strategic partners get access to a curated group of RL researchers, agent founders, technical builders and investors, plus structured opportunities for follow-up. Reach the team at partnership@openstagesai.com.
Where exactly is it?
Sunnyvale, California. The exact address is shared with approved guests before the event.
—

Related reading

December 10, 2026 · Sunnyvale, CA

Agentic AI × RL

Apply on Luma