Deep Dive into MLflow 3.9 Features for Agent Observability and Quality

Description

The MLflow 3.9 release is packed with observability and evaluation features. As part of this webinar, the MLflow team will provide a deep dive into the significant features and the problems they solve. If you are an AI agent developer, an existing MLflow GenAI user, or new to the MLflow GenAI platform, learn how you can incorporate these features as part of your agent development workflow to enhance high-quality agents for production. Some of the features we plan to cover include: 🔹 MLflow Claude Code Assistant 🔹 Dashboard for traces and quality metrics 🔹 Online scheduling for scorers 🔹 Judge building UI 🔹 MemAlign, new optimizer for aligning LLM judge 📅 February 4, 2026 🕒 4:00 PM PT 0:00 Introduction 5:21 MLflow Assistant: Overview and Problem Solved 8:01 MLflow Assistant Demo 12:11 Key Features of MLflow Assistant 14:27 MLflow Assistant Setup and Use Cases 17:41 Upcoming MLflow Assistant Improvements 19:30 Agent Dashboard Metrics: Introduction 22:48 MLflow Agent Dashboard: Solution Overview 27:38 Demo of Agent Dashboard 31:39 Introduction to Judge Builder and MemAlign 37:06 Aligning Judges with Domain Experts using MemAlign 41:26 Continuous Monitoring with Judges 45:03 Upcoming MLflow 3.10 Features 48:50 Q&A Session