Agentic + AI Observability Meetup SF
Join us for Agentic + AI Observability meetup on Thursday. April 9th from 5pm - 8pm PST at the Databricks SF office, an evening focused on agentic architectures and AI observability: how to design, ship, and monitor AI agents that actually work in production.
This meetup is built for engineers, ML 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 reliable agentic systems in the broader open-source ecosystem.
Whether you’re at an early-stage startup or an established company, if you care about getting AI agents into production, and keeping them healthy, this meetup is for you.
Why you should attend
See real architectures: Learn how teams are designing agentic systems on top of data/feature platforms, retrieval, and tools, not just calling a single LLM endpoint.
Learn how to observe what agents are doing: Go beyond logs and dashboards to structured traces, evals, and metrics that help you understand and improve agent behavior over time.
Get hands-on with MLflow and observability tools: Watch live demos of MLflow, tracing integrations, and evaluation workflows for agentic systems.
Connect with other builders: Meet engineers, founders, and practitioners working on similar problems, swap patterns, and find collaborators and potential hires.
Agenda
5:00pm: Registration/Mingling
6:00pm: Welcome Remarks
6:15pm: Talk #1: Building Governed Agents with Databricks (Sunish Sheth, Senior Software Engineer at Databricks)
6:45pm: Talk #2: From Primitives to Production: How Anthropic Builds Agents (Isabella He, Member of Technical Staff at Anthropic)
7:15pm: Mingling with bites + dessert
8:00pm: Night Ends
Session Descriptions:
Building Governed Agents with Databricks
As AI agents move from demos into production, a new class of problem emerges: how do you let agents act autonomously without losing visibility, control, or trust?
This talk dives into the governance challenges that arise when agents call tools, query data, invoke models, and act on behalf of real users — and how Databricks has approached solving them.
From Primitives to Production: How Anthropic Builds Agents
Leaning into model intelligence produces agents that get better as models get better. This session shares what Anthropic has learned building Claude Code and deploying agents with customers across multiple domains and workflows. We break down the core agent loop and the primitives that let the model do more of the work: code execution, skills, MCP servers, hooks, tool permissions, observability, and session management. We'll then show how they compose in a live agent demo of an incident response agent, and discuss how evals can help you continue to iterate on agents in production.