Hands-on AutoResearch: Cracking OpenAI's Parameter Golf
Zhengyao Jiang, Dixing Xu, Vayum Arora, Dhruv Srikanth
- When
- Monday, June 292:20 PM – 4:20 PM · 120 min
- Where
- Track 9San Francisco, CA · imported from ai.engineer's public schedule feed
About this session
Heard about autoresearch, or tried it a few times in playground settings? This hands-on tutorial teaches you how to use autoresearch on one of the most serious challenges in ML this year: OpenAI's Parameter Golf.
The challenge: train the best language model that fits in just 16MB. We entered our autoresearch agent this past spring, and it outperformed the field of over 1,000 participants. You'll learn how we approached it, then get to do it yourself: kick off an autoresearch agent, watch it improve a tiny language model's training script, steer it when progress stalls, and visualize your results. You'll leave with a working autoresearch setup you can point at your own code.
compute kindly sponsored by Modal!
Speakers (4)
CEO & Cofounder, Weco AI
Cofounder & CEO @WecoAI - automated hill climbing with LLMs. Previously: PhD in ML at UCL
Member of Technical Staff, Weco AI
MTS at Weco AI. building self-improving agents and autoresearch systems. Previously at the Linux Foundation.
Growth, Weco AI
Vayum started in AI as an ML engineer and product lead on Apple's Intelligence teams, building the foundational models behind its features. After starting his own company, he backed early-stage founders as a seed investor at Founder Collective, then decided to go build. Now he leads growth at Weco AI, taking recursive self-improvement out of the lab and into ML teams across finance, healthcare, and frontier AI.
Founding Engineer, Weco AI
Dhruv Srikanth is a Founding Engineer at Weco AI, where he works on recursive self-improvement. His work includes AIDE, an autoresearch-style system that achieved nearly 4x the medal rate of the next-best autonomous agent across OpenAI’s MLE-Bench. Previously, Dhruv worked on computer vision and robotics research at Carnegie Mellon University and the University of Chicago.
More in Workshops Day 1
- Cooking with CodexMonday, June 29 · 9:00 AM – 11:00 AM · Track 3
- The best SDLC is the one you build yourself: Why orchestration changes everythingMonday, June 29 · 9:00 AM – 11:00 AM · Track 4
- AI Security Engineer Foundations + CertificateMonday, June 29 · 9:00 AM – 11:00 AM · Track 5
- Total Recall: Agent Memory and Harness EngineeringMonday, June 29 · 9:00 AM – 11:00 AM · Track 6
- Open-Source Inference Engineering for the Agentic EraMonday, June 29 · 9:00 AM – 11:00 AM · Track 8
For developers: this programme is open data — JSON, iCal, schedule XML and an MCP endpoint.Show endpointsHide
- JSONEvery published session and speaker, in one request./aie-worldsfair-2026-import/feed.json
- iCalSubscribe in Google, Apple or Outlook Calendar./aie-worldsfair-2026-import/feed.ics
- Schedule XMLfrab / pentabarf — the format conference apps import./aie-worldsfair-2026-import/feed.xml
- MCP + RESTPoint Claude at the programme. OpenAPI 3.1 included./agents
No key, no signup, CORS open. Everything here is generated from the same data the organisers edit.