As AI systems rapidly move from standalone models to sophisticated multi-component deployments, a new conference will address one of the field’s most critical challenges: how to engineer AI systems that work in the real world. From May 26–29, 2026, researchers and practitioners from academia and industry will gather in San Jose, California for CAIS 2026 , the inaugural ACM Conference on AI and Agentic Systems.
Despite the growing importance of AI systems, there has been no dedicated academic venue focused on how these systems should be designed, optimized, evaluated, and maintained as reliable software artifacts.
“The field at large is still guessing how we can engineer AI software systems that are reliable,” said Omar Khattab, Steering Committee member, Assistant Professor at MIT, and creator of DSPy. “If you want to put a model in a loop, surrounded by control flow and other models, that’s a whole lot harder than talking to ChatGPT yourself. We need to make this an engineering discipline, not a collection of hacks.”
Rather than focusing solely on individual model capabilities, the conference centers on how AI components are composed, orchestrated, and evaluated as part of end-to-end systems.
“Scaling models gets you much of the way, but the hardest problems show up at the system level,” said Matei Zaharia, General Co-Chair, Associate Professor at UC Berkeley, and CTO of Databricks. “Questions of composition, optimization, verification, and evaluation become central when you’re trying to build AI systems that are reliable in practice.”
From Scaffolds to Software
A core theme of CAIS is the difference between language-model scaffolds—temporary structures used to bootstrap new capabilities from today’s models—and AI software systems, which must be durable, efficient, and dependable over time. The conference emphasizes evaluation methods that remain meaningful even as models improve, focusing on application-specific requirements such as latency, correctness, and verifiability.
According to Heather Miller, General Co-Chair, Assistant Professor at Carnegie Mellon University, and Vice President, Research Scientist at Two Sigma Investments, “Compound AI systems require fundamentally different thinking about composition, guarantees, and evaluation. CAIS brings together systems researchers, machine learning researchers, and practitioners to establish shared foundations for this emerging class of software.”
Conference Scope and Program
CAIS 2026 will feature peer-reviewed research across four core areas:
The conference will feature an artifact-centric review process, with eligible submissions receiving ACM reproducibility badges under ACM’s standardized taxonomy, reinforcing CAIS’s commitment to rigorous, reproducible research.
Conference Details
About ACM
ACM, the Association for Computing Machinery, is the world's largest educational and scientific computing society, uniting educators, researchers, and professionals to inspire dialogue, share resources, and address the field's challenges. ACM strengthens the computing profession's collective voice through strong leadership, promotion of the highest standards, and recognition of technical excellence. ACM supports the professional growth of its members by providing opportunities for life-long learning, career development, and professional networking.
###