Anupam Datta

The reality for most enterprises is a "scaling crisis." According to Snowflake's 2026 research with Omdia, 96% of organizations are hitting a wall moving AI from pilot to production. In this landscape, security is no longer just a hurdle — it is the deciding factor between the few who succeed and the majority stuck in experimentation. Securing enterprise agents is one of the biggest pieces of this puzzle. What threats are customers worried about? What's the state of the art in defenses — sandboxing, agent identity and authorization, adversarial testing, security guardrails, observability and audits? Join Snowflake Security Leadership and trailblazing customers for a candid discussion on closing the AI production gap.

Hear from leading faculty from UC Berkeley and Carnegie Mellon about their vision for an agentic future informed by their journeys from fundamental research to widely used open source projects and innovative startups. Learn about the latest innovations with coding agents, data agents, agent safety and the interplay of reasoning models and agents. Hear also about their experiences working with developer communities through their online courses and open source projects.