We are excited to invite submissions for the AI Discovery in the Wild workshop, held in conjunction with the ACM Conference on AI & Agentic Systems (ACM CAIS 2026).

We are bringing together researchers and practitioners building autonomous AI systems that search, optimize, and discover. We are particularly focused on the hard problems of taking these agents out of the lab and into the real world—specifically in science, engineering, and systems infrastructure where validation is difficult and human oversight matters.

🚨 Concurrent Submissions Highly Encouraged
With the NeurIPS deadline right around the corner, we want to make submitting as frictionless as possible. Because our workshop is strictly non-archival, we highly encourage you to submit the concurrent drafts for NeurIPS or other main-track conferences!

Invited Speakers:
- Aditya Akella (UT Austin)
- Mohammad Alizadeh (MIT / Glia AI)
- Joseph E. Gonzalez (UC Berkeley)
- Azalia Mirhoseini (Stanford / Ricursive Intelligence)
- Graham Neubig (CMU / OpenHands)
- Ion Stoica (UC Berkeley)
- James Zou (Stanford)
- Martin Maas (Google)

Topics of interest include (but are not limited to):
- Search and optimization with LLM agents
- Agents for systems infrastructure
- Methods for evaluation under real-world constraints
- Deployment reports and failure analysis
- Human-agent collaboration & oversight

Submission Details:
- Format: 4-page short papers or 9-page long papers.
- Submission Site: Available via our website
- Deadline: May 4, 2026 (AoE)

Organizing Committee:
Shubham Agarwal (UCB), Lakshya Agrawal (UCB), Mert Cemri (UCB), Alex Dimakis (UCB/Bespoke Labs), Batu El (Stanford), Alex Krentsel (Google/UCB), Eric Liang (Databricks), Shu Liu (UCB), Rui Meng (Google), Sylvia Ratnasamy (UCB), Ion Stoica (UCB), and Matei Zaharia (UCB/Databricks).

Feel free to reach out to the organizing team with any questions—we look forward to seeing your work in San Jose!

Contact: cais26.ai.discovery@gmail.com