Dmitri Ryssev

In this session, discover how Dataiku extends the Databricks lakehouse to accelerate AI and agent development, orchestration, and governance. We will explore how teams can rapidly build AI projects with Cobuild, Dataiku's embedded conversational agent; create and orchestrate intelligent agents leveraging Databricks capabilities like Vector Search and Genie; and ensure enterprise-grade governance with built-in lifecycle management, monitoring, and controls. Learn how organizations are moving from experimentation to production faster by combining Databricks' powerful data platform with Dataiku's unified AI layer, enabling seamless collaboration across data, engineering, and business teams to deliver real-world AI at scale.

AI agent systems hold immense promise for automating complex tasks and driving intelligent decision‑making, but only when they are engineered to be both resilient and transparent. In this session we will explore how Dataiku’s LLM Mesh pairs with Databricks Mosaic AI to streamline the entire lifecycle: ingesting and preparing data in the Lakehouse, prompt engineering LLMs hosted on Mosaic AI Model Serving Endpoints, visually orchestrating multi‑step chains, and monitoring them in real time. We’ll walk through a live demo of a Dataiku flow that connects to a Databricks hosted model, adds automated validation, lineage, and human‑in‑the‑loop review, then exposes the agent via Dataiku's Agent Connect interface. You’ll leave with actionable patterns for setting guardrails, logging decisions, and surfacing explanations—so your organization can deploy trustworthy domain‑specific agents faster & safer.