​Join us for Agentic + AI Observability meetup on Tuesday, September 29 from 5pm - 8pm PST at the Databricks Bellevue office, for an evening focused on agentic architectures and AI observability: real-world deployment, smart routing, and automated incident resolution.

​This meetup is built for engineers, ML/AI practitioners, and AI startup founders who are already experimenting with agents (or planning to) and want to go deeper into the tech. We’ll cover real-world patterns, failure modes, and tooling for building a reliable meta-harness layer to manage agent execution, control and choice, state, and runtime policies across the broader ecosystem.

​Whether you’re at an early-stage startup or an established company, if you care about getting AI agents into production and keeping multi-agent systems healthy, this meetup is for you.

​Previous series speakers include Databricks, Anthropic, OpenAI, SpaceXAI and more.

​Why you should attend

• ​See real agentic architectures: Learn how teams design smart routing and meta-harness layers using Unity Gateway to optimize model execution without sacrificing caching performance.

• ​Explore automated incident response & tracing: See how tools like Genie ZeroOps handle structured root-cause investigations, evaluate agent choices, and automate system recovery when workloads fail.

• ​Connect with other builders: Meet engineers, founders, and practitioners working on similar problems, swap design patterns, and expand your network.

​Agenda

• ​5:00pm: Registration/Mingling

• ​6:00pm: Welcome Remarks

• ​6:15pm: Talk #1: Route The Task, Not The Request: Smart Routing Without Breaking The Cache (Elise Gonzales, Staff Product Manager at Databricks)

• ​6:45pm: Talk #2: Investigating & Resolving Data and AI incidents with Genie ZeroOps (Harshit Rai, Sr. Forward Deployed Engineer at Databricks)

• ​7:15pm: Mingling with bites + dessert

• ​8:00pm: Night Ends

​Session Descriptions:

• ​Route The Task, Not The Request: Smart Routing Without Breaking The Cache

  • ​Routing sounds simple: send easy tasks to cheap models, hard tasks to expensive ones. In practice the hard parts are deciding what "hard" means, choosing where in the execution flow to decide, and not destroying prompt caching in the process. This talk walks through the design of Unity Gateway Smart Routing, which selects both the model and the agent harness for each coding task. We'll cover the classification signals, the decision boundaries (session start and subagent launch, not per request), why harness and model have to be chosen together, and what's coming next

• ​Investigating & Resolving Data and AI incidents with Genie ZeroOps

  • ​It’s 2:00 a.m. A critical data pipeline fails, dashboards go stale, and the on-call engineer starts digging through job runs, logs, dependencies, and recent changes. What if Genie could take the first shift? This talk walks through Genie ZeroOps, an emerging Databricks capability for monitoring data and AI workloads, investigating failures and data-quality anomalies, identifying likely root causes, and proposing repairs. We’ll follow an overnight incident from the initial alert to impact analysis, safe validation in an isolated environment, and final human approval—showing how Genie can take on investigation and repetitive toil without taking control away from the engineer.