This track brings together the team's shipping state‑of‑the‑art systems to unpack how modern AI moves from prototype to production. Across sessions, you’ll see how code-first copilots are built and evaluated (Anthropic), and how reinforcement learning optimizes large‑scale marketplaces and ads (Meta). We’ll dive into AI content generation and agent‑based simulation for creative and safety‑aware workflows (Pinterest + Veris AI), personalization with LLMs that balances latency, cost, and relevance (DoorDash), and the emerging Model Context Protocol for interoperable tooling and orchestration (Browserbase). 

Expect concrete patterns for retrieval and data pipelines, multi‑agent design, online/offline evals, guardrails and governance, observability, and continuous improvement loops, so technical leaders can deliver scalable systems that are reliable, compliant, and measurably impactful.