Loading…
AI DevSummit 2025 + DeveloperWeek Leadership 2025
← Back to schedule
OPEN Session: Evolving Databricks’ Kubernetes Infrastructure to Scale AI Workloads via Principled Platform Abstrac
Wednesday May 28, 2025 2:00pm - 2:25pm PDT
DeveloperWeek Leadership Expo Stage
Sourav Khandelwal, Databricks, Software Engineer

In the age of large-scale AI—where workloads like LLM model serving and Retrieval-Augmented Generation (RAG) demand massive compute capacity on short notice—efficient, reliable, and self-service infrastructure is essential. At Databricks, we faced the challenge of orchestrating a rapidly growing fleet of Kubernetes clusters spanning multiple clouds, regions, and use cases to power AI-driven solutions. The result was a patchwork of manual processes that were both costly in engineering hours and prone to human error.

In this session, you’ll learn how at Databricks, we overcame these challenges by building principled platform abstractions—clean, self-serve interfaces that application teams can use to provision, configure, and manage their Kubernetes clusters in a standardized, automated way. We’ll detail our journey from fragmented workflows to a scalable infrastructure platform that underpins products such as LLM model serving and RAG, allowing for quick, repeatable scale-out of AI-driven workloads.

We’ll also share the hard lessons learned, including the trade-offs between flexibility and uniformity, as well as strategies for ensuring consistent cluster management patterns across diverse environments. By highlighting how these platform abstractions free up engineering resources and streamline high-demand AI use cases, we’ll showcase how this approach accelerates product development cycles while simplifying day-to-day operations.
Speakers
avatar for Sourav Khandelwal

Sourav Khandelwal

Software Engineer, Databricks
I am a seasoned software engineer with over 12 years of experience in designing and managing large-scale platforms in cloud-native environments. At Databricks, I have led and contributed to several innovative projects that have scaled and automated our Kubernetes Compute Platform... Read More →
Wednesday May 28, 2025 2:00pm - 2:25pm PDT
DeveloperWeek Leadership Expo Stage
  AI DevSummit
Log in to leave feedback.

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Tweet Share
Need help? View Support Guides
Event questions? Contact Event Planner
©2025 Sched • About • Privacy • Terms
Share Modal

Share this link via

Or copy link