AI Council 2027: March 23–24 at the Palace Hotel

The technical AI conference bringing together engineers, researchers, and founders in San Francisco to explore the infrastructure, systems, and engineering behind production AI.

AI Council returns on March 23–24, 2027, at San Francisco's historic Palace Hotel, bringing together the people building the systems that power AI.

Organized by Zero Prime, AI Council is a community-led technical conference focused on the engineering challenges behind bringing AI into production. Over two days, attendees will explore real architectures, infrastructure decisions, and lessons from the engineers and researchers working at the forefront of the field.

The program includes technical talks, hands-on workshops, speaker office hours, and opportunities to connect directly with the people building the AI stack.

A decade of bringing technical builders together

AI Council began as Data Council in 2015, bringing together the engineers and architects building modern data infrastructure. Over the past decade, that community has evolved alongside the technology, from distributed databases and data platforms to the systems now powering AI.

The name has changed, but the approach hasn't.

AI Council is built around technical depth, not product pitches. Talks are curated by practitioners working in the field, with an emphasis on real systems, architectural decisions, performance tradeoffs, and lessons learned in production.

The goal is to create a place where engineers can learn directly from each other, challenge assumptions, and discuss the difficult problems that rarely make it into polished product demos.

Learn more about AI Council.

From AI Council 2026: The people behind the technology

The 2026 conference featured engineers, researchers, and founders sharing work that is shaping the direction of AI.

Diogo Almeida, a former OpenAI researcher who helped develop ChatGPT, presented AI: Too Good to Be True, Too Bad to Be Useful, examining why language models that excel at assisting people can struggle with reliable automation. Just months later, his company, TypeSafe AI, emerged from stealth with $40 million in funding and introduced Jev, a model designed for software-driven decisions.

Hannes Mühleisen, co-creator of DuckDB, used his session, Super-Secret Next Big Thing for DuckDB, to unveil Quack, a new protocol extending DuckDB's capabilities beyond its traditional embedded architecture.

And Linus Lee, Head of AI at Thrive Capital, explored the practical challenges of building capable AI assistants in Context Engineering at the Frontier, covering retrieval, tool use, structured data, and the challenge of finding useful signals in increasingly complex systems.

These talks reflect what makes AI Council different: the people developing the technology share not just what they're building, but how it works and the problems they're trying to solve.

Explore talks from AI Council 2026.

Eight technical tracks spanning the AI stack

AI is moving beyond experimentation, creating new challenges across infrastructure, models, agents, and applications.

The 2027 conference will feature eight technical tracks covering the systems and engineering disciplines shaping this transition.

• AI Engineering: Frameworks, tools, and engineering practices for building AI applications.
• AI Security & Safety: Securing models, agents, data, and AI-powered systems.
• Agent Infrastructure: Runtime environments, orchestration, memory, and infrastructure for autonomous agents.
• Software Factories: Coding agents, automated development workflows, and the changing practice of software engineering.
• Model Systems: Inference, optimization, serving, and large-scale model deployment.
• Data Infra for AI: Databases, retrieval, storage, and data pipelines supporting AI workloads.
• Agents in Production: Reliability, evaluation, observability, and lessons from operating agents at scale.
• Physical AI & Robotics: Embodied intelligence, robotics, and the systems connecting AI to the physical world.

Across every track, the focus is on practical engineering knowledge and the decisions that matter when systems move into production.

Call for speakers: Show us what you've built

Speaker applications for AI Council 2027 are now open.

We're looking for engineers, researchers, and technical founders who can share firsthand experience building or operating AI systems.

The strongest talks go beyond a product overview. They explain a specific technical challenge, the architecture behind a solution, the tradeoffs involved, and what worked or failed along the way.

You don't need a previous speaking record. We care about the technical work and what others can learn from it.

Applications close December 11, 2026.

Apply to speak at AI Council 2027.

Beyond the stage

AI Council is designed for conversations that go deeper than a conference presentation.

Speaker office hours give attendees an opportunity to ask detailed technical questions and continue discussions after each session. Hands-on workshops provide space to explore tools and techniques in practice.

The conference also brings together infrastructure companies, emerging startups, and engineering teams working across the AI ecosystem.

From informal conversations to community events around San Francisco, AI Council creates opportunities to connect with people solving similar problems at different layers of the stack.

Join us at the Palace Hotel

AI Council 2027 takes place March 23–24 at the Palace Hotel in San Francisco.

Whether you're designing AI infrastructure, deploying agents, developing model systems, or bringing AI applications into production, AI Council is a place to learn from the people doing the work.

Explore AI Council 2027 and register.

‍