Platform Engineering & AI

Join us for an evening focused on the tools and cultural shifts driving the industry forward. We’ll dive into AI, Platform Engineering, and Open Source. With practical demos, talks, and time to connect with other builders. Speakers: Allison Ding | Developer Advocate Data Science @ NVIDIA Sriharsha Yayi | Staff Product Manager @ Intuit Balachandra Maddina | Principal SWE @ Intuit Steven Liu | Technical Writer Developer Docs @ HuggingFace Agenda: 4:30 PM — Check-in opens 5:25 PM — Welcome + kickoff 5:30–5:55 PM — Tech Talk 1 6:00–6:25 PM — Tech Talk 2 6:30–6:55 PM — Tech Talk 3 7:00–8:00 PM — Networking + happy hour Tech Talk 1: Prompts to Pixels with Diffusers Speakers: Steven Liu, HuggingFace About this talk: This talk introduces the Diffusers library, an open-source toolkit for running and customizing diffusion models like Flux. Learn how diffusion models start from pure noise and transform it into images or videos and why prompting matters. We'll end with a live demo on Hugging Face Spaces where we prompt a model together and generate images or input images and generate other images. Tech Talk 2: Build Production-Ready Async AI Agents with Numaflow Speakers: Sriharsha Yayi and Balachandra Maddina, Intuit About this talk: We will use open-source Numaproj to build asynchronous AI Agents. Tackling one of our biggest Accounting Agent applications, learn how event-driven architecture decouples high-latency LLM calls to prevent system blocking. We'll end with a live demo where we set up a pipeline, send an event to a Kafka topic, and watch responses in the logs. Tech Talk 3: End-to-End GPU-Accelerated Data Science Speaker: Allison Ding, NVIDIA About this talk: This talk introduces NVIDIA's ecosystem of GPU-accelerated data science (for data processing, machine learning) — cuDF, Polars, cuML, and Triton — and shows how they can dramatically speed up every stage of a data science workflow with minimal code changes from familiar Python APIs. Using a real-world fraud detection dataset, we'll walk through live benchmarks at each stage and finish with deploying a model to production using Triton for millisecond-latency inference. We'll also demo an end-to-end ML agent powered by vLLM inference, running on DGX Spark. Come meet fellow engineers and builders, explore what’s being built at Intuit and beyond, and connect with the open-source community. Hope to see you there! About Intuit Open Source: Intuit is dedicated to open source, not only as users but also by maintaining and contributing to various tools and communities like CNCF, Argo, Numaproj, Istio, GraphQL, Kubernetes, and web/mobile ecosystems. Join our LinkedIn open source community for updates, best practices, and discussions. Please note this event is restricted to individuals aged 21 and above due to venue regulations.