Building the Future of Agentic & AI Systems
ACM CAIS 2026 — The premier venue for rigorous, reproducible research on compound AI architectures, optimization, and deployment.
CAIS hotel room block & rates available until April 26 May 15
DoubleTree by Hilton San Jose
2050 Gateway Place, San Jose, CA 95110
Complimentary shuttle to/from San Jose International Airport (SJC) • Discounted parking $10/day
Group Rate: $159/night + tax
Book HotelKeynote Speakers
Hear from leaders shaping the future of AI and agentic systems.
Andy Konwinski
Co-founder of Databricks and Perplexity AI · Founder of Laude Institute
Laude's "Ship Your Research" mission funds open-source AI research through its Moonshots, Slingshots, and Open Frontier programs. Terminal-Bench, a Laude-backed agent benchmark, has become an industry-wide standard for measuring command line performance.
Thariq Shihipar
Member of Technical Staff, Claude Code · Anthropic
Thariq is a core builder of Claude Code, the agentic coding tool that has become one of the most widely adopted developer interfaces for working with LLMs. His technical writing on prompt caching, tool design, and "unhobbling" has shaped how practitioners think about building reliable agentic systems.
Percy Liang
Professor at Stanford · Co-founder of Together AI and Simile AI · Creator of Marin
Percy Liang is a Professor of Computer Science at Stanford University, co-founder of Together AI and Simile AI, and the creator of Marin, which aims to build frontier models fully in the open. He has made a number of contributions in AI, including the SQuAD question answering dataset, the HELM benchmarking framework, generative agents, prefix tuning, and coining the term "foundation models".
The 2026 Program
Four days in San Jose, May 26–29. One day of workshops and tutorials, followed by three days of peer-reviewed research papers, system demonstrations, and keynotes.
Tue, May 26
5 Workshops & Tutorials
Day-long programs on agent skills, agentic software engineering, RL environments, discovery agents, and more.
Browse WorkshopsWed–Fri, May 27–29
63 Research Papers
Peer-reviewed contributions on architecture, optimization, evaluation, security, and engineering of AI and agentic systems.
Browse PapersWed–Fri, May 27–29
46 System Demos
Working implementations of AI systems and agent systems, presented live by their authors.
Browse DemosParticipating Institutions
CAIS 2026 papers and demos come from 115+ institutions across academia and industry, including:
At a Glance
| Tue, May 26 | Workshops & Tutorials |
| Wed, May 27 | Main Conference — Day 1 (papers, demos, keynote) |
| Thu, May 28 | Main Conference — Day 2 (papers, demos, keynote) |
| Fri, May 29 | Main Conference — Day 3 (papers, demos, keynote) |
CAIS × AI Engineer World's Fair
ACM CAIS 2026 has partnered with the AI Engineer World's Fair (June 29–July 2, Moscone West, San Francisco). Accepted CAIS papers that earn an Industry Spotlight or Operational Experience designation will be invited to present at both venues — giving authors peer review through ACM plus a stage in front of 6,000+ practicing engineers.
What is ACM CAIS?
A high-signal forum for rigorous, reproducible research on agentic and AI systems—architectures that shift the Pareto frontier through principled composition of multiple system components, smart inference-time scaling strategies, and systematic verification methods for reliable deployment.
Architectural Patterns & Composition
Networks of Networks and inference-time scaling architectures. Verifier-based systems leveraging generation/verification asymmetry. RAG, multi-agent, and tool-augmented designs.
System Optimization & Efficiency
End-to-end optimization of non-differentiable pipelines. Cost-performance trade-offs, resource allocation across components, and automated architecture search for compound systems at scale.
Engineering & Operations
MLOps for compound AI. Monitoring, debugging, observability, and specifications. Security and safety in multi-component systems. Production deployment case studies.
Evaluation & Benchmarking
Metrics for compound system performance. Reproducibility frameworks and artifact standards. Comparative evaluation methodologies and real-world impact assessment.
Conference Leadership
ACM CAIS is guided by leading researchers who defined the compound AI systems paradigm—scholars and practitioners from MIT, Stanford, Google DeepMind, Microsoft Research, and beyond.
Steering Committee
Organizing Committee
Be part of the inaugural ACM CAIS
Join us May 26–29, 2026 in San Jose. Registration is open.
Become a Sponsor
Put your organization in front of 500+ researchers and practitioners defining the future of compound AI systems — from top university labs to industry teams deploying agentic architectures at scale.
Gold $20,000 · Silver $12,000 · Bronze $6,000 · À la carte from $1,500