Platform Engineer, AI Infrastructure at Sarvam AI (Bengaluru)

Job Description

Sarvam AI is hiring a Platform Engineer for AI infrastructure in Bengaluru to build the scheduling, scaling, multi-tenancy and serving layers that let its ML teams use thousands of GPUs, shared between training jobs spanning hundreds of GPUs and latency-sensitive inference, without a person in the loop for every job.

At a glance

• Role: Platform Engineer, AI Infrastructure
• Location: Bengaluru
• Experience: 5+ years building infrastructure or platform software
• Core skills: Kubernetes at controller level, Go or Python, GPU scheduling
• Apply: open until filled (re-checked 12 October 2026)

What you could build

• The serving platform: turning model artifacts into scalable multi-tenant endpoints with routing, load balancing, canary and blue-green rollouts and traffic splitting
• Scaling and elasticity: autoscaling for training (elastic and gang scaling) and serving, capacity pooling, preemption and GPU bin-packing
• Scheduling with Kueue, Volcano, Slurm-on-Kubernetes or custom controllers, including gang scheduling, quotas, fairness and topology-aware placement
• Multi-tenancy and isolation: tenants, RBAC, quotas, and MIG, MPS or time-slicing offered as self-service tiers
• Networking, observability and cost tooling, and storage abstractions over the parallel filesystem
• Developer tools (CLI, SDK, APIs) and infrastructure-as-code for reproducible multi-vendor cluster bring-up

The posting draws a clear line: Sarvam’s SREs keep the fleet reliable and carry the pager, while this role builds the software that makes the fleet usable and less fragile. It is heavy software engineering that treats the platform as a product and ML engineers as its customers, and you will take one capability at a time and build it end to end.

What Sarvam AI is looking for

• 5+ years building infrastructure or platform software, with services and control planes others built on
• Strong Go or Python and the ability to debug systems in production
• Kubernetes at the controller and internals level, including writing operators or controllers
• Working knowledge of GPU platform constraints: MIG and GPU sharing, gang scheduling, topology-aware placement, and why training and serving compete for hardware
• A product mindset toward internal users, and the range to own a capability from design to documentation

Nice to have

• Having built a serving, inference or training platform
• GPU schedulers such as Kueue, Volcano, Slurm or Run:ai in multi-tenant production
• Multi-tenant GPU isolation shipped as a self-service capability
• Deep Kubernetes networking (CNI internals, RDMA or SR-IOV in pods)
• On-premise GPU platform work, including multi-vendor or Indian NCP environments
• Open-source contributions to Kubernetes, scheduling or GPU projects

Pay and location

Sarvam’s posting does not state a salary. The role is based in Bengaluru.

How to apply

Apply through the official Sarvam AI job posting. Sarvam AI’s posting does not give a closing date, so the role is open until filled; ResearchJobs.in will re-check this listing on 12 October 2026. Check the posting for location and work-authorisation details before you apply.

See all our artificial intelligence jobs, or browse more research jobs on ResearchJobs.in.