Scalable and Resilient Agentic Systems

Workshop at ICML 2026

When and Where: The workshop will be held on July 10th or 11th, 2026 (one day), following the ICML main conference at the COEX Convention & Exhibition Center in Seoul, South Korea. This is an in-person workshop.

Scope: This workshop addresses the critical system-level challenges in building scalable, robust, and reliably deployable agentic systems. As AI agents become more sophisticated and interconnected, understanding how to design systems that can scale efficiently while maintaining resilience becomes paramount.

Join us to explore and contribute to the future of robust agentic systems!

🤖

Workshop at

ICML 2026

Scalable & Resilient
Agentic Systems

July 10-11 • Seoul 🇰🇷

Call for Papers

Topics: We welcome and highlight content on scalable and resilient agentic systems, including but not limited to:

•
Scalability: Scaling multi-agent systems; distributed coordination; efficient resource use in large deployments.
•
Training Efficiency: Sample and compute efficient fine-tuning for agentic behaviour; RLHF; curriculum and continual learning for agents.
•
Multi-Agent Coordination: Task allocation; communication protocols; emergent behaviors.
•
Novel Architectures: Diffusion language models; multi-modal agents; new LLM architectures for agentic systems.
•
Robustness: Handling failures; recovery mechanisms; testing agentic systems; learning coordination strategies across heterogeneous agents.
•
Safety: Safety constraints; alignment in multi-agent settings; monitoring and oversight.
•
Security: Attacks (such as adversarial inputs and environments) and defenses for agentic and multi-agent systems.
•
Privacy: Privacy-preserving training and deployment of agentic systems; privacy attacks on existing systems.
•
Real-world Deployment: Production systems; real-world case studies; continuous and online learning.
•
Evaluation: Metrics and benchmarks for agentic systems at scale; evaluating generalization across environments; measuring training efficiency for agentic capabilities.