Introduction to GPU Infrastructure

#art #computer-science #learning #machine-learning #ieeeday #data-science

Grace Gong and Ganeshkumar Ashokavardhanan will be presenting on "Intro to GPU infrastructure", covering the systems behind production AI workloads, including GPU compute, networking, orchestration, and reliability.

Grace Gong
Grace Gong is an engineer at Microsoft focused on AIOps and reliability across Azure. She previously worked on the Microsoft Fabric data platform as a product manager and on machine learning and large language model observability at Datadog. She is finishing up her master’s in computer science at Georgia Tech and is part of the founding NVIDIA Developer Champions cohort. Grace enjoys making technical concepts accessible through talks, mentoring, and hands-on learning. In this IEEE Seattle workshop, she’ll introduce GPU fundamentals and explain how GPUs support modern AI workloads.

Grace Gong | LinkedIn

Ganeshkumar Ashokavardhanan
Ganeshkumar Ashokavardhanan is a Senior Software Engineer at Microsoft and the Tech Lead for GPU Infrastructure on the Azure Kubernetes Service (AKS) team, where he works on making GPUs easy to use, reliable, and configurable on Kubernetes. He led the fully managed GPU experience on AKS, which makes provisioning a GPU node as simple as a general-purpose one, and GPU health monitoring, which automatically detects and reports issues affecting GPUs. He also shipped Artifact Streaming on AKS, which significantly reduces startup times for workloads with large container images. Ganeshkumar has given six KubeCon talks across North America, Europe, and Hong Kong on GPU infrastructure, AI agents, and faster workload startup. He also presented research on evaluating AI agents that act on live systems at ACM CAIS 2026, and holds dual degrees in EECS and Business from UC Berkeley's M.E.T. program. In this IEEE Seattle workshop, he'll cover GPU infrastructure from a practitioner's perspective, drawing on his experience shipping features that companies use to run GPU fleets in production.

Ganeshkumar Ashokavardhanan | LinkedIn