SF AI Show and Tell: July 2025

Catch up on demand! This installment of SF AI Show and Tell originally took place on July 1 and featured talks by Philip Kiely, Michael N Magan, and SallyAnn DeLucia.

Talks: Philip Kiely @ Baseten - Five things to know about AI infrastructure as you scale

​In 2025, AI-native startups are scaling to 8 figures of annual revenue in months. How do you build infrastructure that keeps up with this explosive growth? This talk provides five sometimes-surprising solutions selected from the hardest challenges we’ve faced while working on scaling fast-growing AI startups.

​Michael N Magan @ Tambo.co - Generative UI in Practice: From Intent to Interface

​Generative UI lets AI assistants surface real UI such as charts, forms, and complete workflows exactly when users need them. This talk shows how it works by treating components as tools: mapping user intent to the right component, binding data, and rendering on the fly. A demo of how we use it internally. I'll close with the main opportunities, and the chief challenges. Attendees leave knowing what generative UI is, why it matters, and how to ship it.

​SallyAnn DeLucia @ Arize - Evaluating Agents in the Wild with Arize

​This talk will demo how Arize supports agent evaluation with a focus on strategy and visibility. I’ll walk through our new Agent View, which aggregates traces, tool use, and reasoning steps into a structured, searchable format—making it easier to debug failures, track performance, and iterate on agent design.

[eventID:24476]