Infrastructure SRE - HPC

The role

This person will operate a large multi-vendor GPU fleet for training jobs and inference services. They will manage provisioning, observability, capacity, and fleet health while building internal tooling.

What you will do

• Operate the GPU fleet end to end across training and serving
• Provision, observe, and manage capacity and fleet health
• Hold a meaningful on-call rotation
• Write runbooks that hold up under pressure
• Drive postmortems that produce durable fixes
• Build internal tooling
• Partner with ML and platform teams to keep large runs alive and serving latency predictable

Required qualifications

• 5+ years in infrastructure or site reliability engineering
• 2+ years operating GPU clusters at scale
• Demonstrated on-call ownership of infrastructure
• Track record of postmortems that led to real change
• Proficiency in Python or Go
• Working fluency across all five areas of focus

Nice to have

• Slurm and Kubernetes hybrid environments
• On-premise GPU deployment, including coordination with datacenter operations on power, cooling, and InfiniBand cabling
• Experience with Indian NCPs, DGX SuperPOD, Lambda, CoreWeave, NeevCloud etc.
• Multi-tenant GPU isolation (MIG, MPS, time-slicing) in production

About the company

Sarvam is building the bedrock of Sovereign AI for India by developing a full-stack sovereign AI platform. The company works with leading enterprises and public institutions and is backed by Lightspeed, Peak XV, and Khosla Ventures.