---
title: 'ML Engineer (Training Infra), Foundational Models at Sarvam'
canonical: 'https://feeny.ai/job/ml-engineer-training-infra-foundational-models-sarvam-bengaluru-byw6f3s2ty06'
type: 'job'
last_seen: '2026-10-17'
---

# ML Engineer (Training Infra), Foundational Models at Sarvam

- **Company:** [Sarvam](https://feeny.ai/companies/sarvam)
- **Location:** Bengaluru, India
- **Employment:** full-time
- **Work type:** onsite
- **Posted:** 2026-10-07
- **Last confirmed live:** 2026-10-17
- **Apply:** https://jobs.ashbyhq.com/sarvam/e9fb1269-fc4b-4879-aa35-0b75a8c83611/application

**Skills:** Distributed training frameworks, Megatron-LM, DeepSpeed, FSDP, NeMo, CUDA, Triton, PyTorch, Nsight, PyTorch profiler, GPU architecture, Python, Custom CUDA kernel development, Triton kernel development, Slurm, Kubernetes, Mixed precision training, Quantization-aware training, BF16, FP8

> Own the infrastructure for training foundational models, focusing on distributed training at scale, parallelism strategies, and reliability engineering. The role involves optimizing throughput, tuning GPU kernels, and ensuring fault tolerance for large-scale AI model training.

## Job description

## About Sarvam

Sarvam is building the bedrock of Sovereign AI for India. The company is developing India’s full-stack sovereign AI platform, building across research, models, infrastructure and applications with a singular focus on making AI genuinely work for India. Sarvam works with leading enterprises and public institutions and is backed by Lightspeed, Peak XV, and Khosla Ventures. Sarvam partners with India’s leading brands, including Tata Capital, SBI Life, CRED, IDFC, and LIC.

## About the Role

You will own the infrastructure that our next family of foundational models is trained on. This is a deep systems role: distributed training at scale, parallelism strategies, GPU kernel work, throughput and stability optimization, and the unglamorous reliability engineering that determines whether a frontier training run actually finishes. The bar is high. A bad MFU number, a slow data loader, a switch flap — these are problems whose cost is measured in weeks and millions of dollars. You will be expected to find them before they happen, and to fix them fast when they do.

## What You’ll Do

- Build, maintain, and continually push the limits of our distributed training stack across large GPU clusters.
- Design and implement parallelism strategies — data, tensor, pipeline, sequence, expert — and reason about which combinations make sense for which architectures and which scales.
- Profile and optimize end-to-end training throughput: kernel performance, communication overlap, memory layout, checkpointing, data loading.
- Write and tune custom GPU kernels (CUDA, Triton) where off-the-shelf is leaving performance on the table.
- Own the reliability of long-running training jobs — fault tolerance, checkpoint integrity, deterministic restarts, automated detection of slow nodes and silent corruption.
- Partner closely with researchers to make sure architectural ideas can actually be trained efficiently, and with the data team to keep the pipeline from being the bottleneck.

## What We’re Looking For

- BS or MS in Computer Science or a closely related technical field (or equivalent demonstrated experience).
- 3+ years of experience building ML training infrastructure or large-scale distributed systems. Exceptional early-career candidates with a strong systems background will be considered.
- Hands-on experience training large models with distributed training frameworks — Megatron-LM, DeepSpeed, FSDP, NeMo, or equivalent. You should have been on-call for a real pretraining run.
- Deep working knowledge of GPU architecture, the CUDA programming model, and standard tools for profiling GPU workloads (Nsight, PyTorch profiler, etc.).
- Strong PyTorch internals knowledge. You're comfortable reading and modifying low-level training code, not just using high-level APIs.
- Meaningful open-source contributions in the training infrastructure ecosystem — Megatron, DeepSpeed, PyTorch, vLLM, Triton, NCCL, or similar.

Bonus Points

- Custom CUDA or Triton kernel development with measurable wins on real workloads.
- Direct experience training models at 10B+ parameters or on 1000+ GPU clusters.
- Familiarity with cluster orchestration and job scheduling (Slurm, Kubernetes) at scale.
- Experience with mixed precision (BF16, FP8), quantization-aware training, or other numerical work that has shown up in production training runs.
- First-author papers or technical reports on training systems, scaling, or model efficiency.

Why this role? Frontier model training is one of the hardest systems problems in the world right now, and most of the people doing it well sit at a small handful of labs. You will be doing the same work, with the same autonomy, and your impact will be visible in every model Sarvam ships. Why Sarvam? Sarvam is a fast-moving, high talent-density team building full-stack AI for India, working on problems that push the frontiers of AI with real population-scale impact.

- Work alongside researchers, engineers, builders, and business leaders who move fast and hold each other to a very high bar
- High ownership and high impact, from day one
- Everything we do is AI-first, from the way we build and ship to the way we think about problems
- You can work on problems that could change how an entire country learns, works, and communicates

If you want to work on problems at the frontier of AI in India, Sarvam is the place to be.

## About Sarvam

## Company Overview
- **One-liner**: Sarvam is building India's full-stack sovereign AI platform, delivering state-of-the-art models and enterprise-grade applications for Indic languages.
- **Entity Type**: Private (Series A – $41M raised)
- **Headquarters**: Bengaluru, India
- **Founded**: August 2023
- **Founders**: Dr. Vivek Raghavan, Dr. Pratyush Kumar

## Core Business
- Primary industry: Artificial Intelligence / Generative AI
- Target customers: Enterprises, Government, Developers (B2B, B2G)
- Mission: Make AI widely accessible to everyone in India by building foundational components that reflect how the country thinks, speaks, reasons, and solves problems.

## Products & Services
- **Sarvam Cloud**: Fully managed, automatically scaling API platform for Indic AI capabilities.
- **Speech to Text (Saaras)**: High‑accuracy ASR across 12 Indic languages.
- **Text to Speech (Bulbul)**: Natural voices across 11 Indic languages.
- **Translation**: State‑of‑the‑art translation across 23 languages.
- **Document Digitisation (Akshar)**: Convert PDFs and images into structured data.
- **Vision Models**: Image understanding and processing.
- **Samvaad**: Conversational AI studio for building voice agents.
- **Arya**: AI‑for‑work platform with forward‑deployed engineers for custom agent design and integration.
- **Deployment Options**: Private cloud, on‑premise, hybrid, or air‑gapped environments.

## Market Standing
- **Valuation/Market Cap**: Not disclosed (private)
- **Key Metric**: Total funding $41M (from Khosla Ventures, Lightspeed, Peak XV Partners)
- **Notable Investors/Partners**: Khosla Ventures, Lightspeed, Peak XV Partners
- **Growth Signals**: 62 open positions across engineering, models, product, sales, and more; case studies with large‑scale deployments (50,000+ farmer feedback calls, millions of policy calls); SOC 2 Type II, ISO 27001, DPDP compliant.

## Competitive Advantages
- **Sovereign AI**: All technology developed and operated entirely in India, with full data control and compliance with Indian data protection laws.
- **Full‑stack platform**: From foundational models to enterprise agents, everything built in‑house.
- **Forward‑deployed engineering**: Dedicated engineers embedded with customer teams to ensure successful delivery.
- **Population‑scale impact**: Proven ability to serve millions of users in Indic languages across government and BFSI.
- **Deployment flexibility**: Can run on any infrastructure, including air‑gapped for regulated industries.

## Strategic Focus
- Continue building state‑of‑the‑art models and tools for Indic languages.
- Expand enterprise and government adoption of conversational AI agents.
- Nurture the Indian AI ecosystem by open‑sourcing models and tools.
- Scale team rapidly (62 open roles) to support growing customer demand.

## Why Work Here
- **Mission‑driven opportunity**: Shape AI that serves India’s linguistic and cultural diversity.
- **On‑site first**: All roles are based in Bengaluru (Delhi also available), fostering deep collaboration.
- **Cutting‑edge work**: Foundational model research, ML infrastructure, full‑stack product engineering.
- **High‑impact culture**: Teams work directly with customers via forward‑deployed engineers, seeing immediate real‑world use.
- **Backed by top investors**: Financial stability and growth runway from Lightspeed, Peak XV, and Khosla Ventures.

## Sources
1. [sarvam.ai](https://www.sarvam.ai/)
2. [sarvam.ai/about-us](https://www.sarvam.ai/about-us)
3. [sarvam.ai/careers](https://www.sarvam.ai/careers)
4. [jobs.ashbyhq.com/sarvam](https://jobs.ashbyhq.com/sarvam)
5. [linkedin.com/company/sarvam-ai](https://www.linkedin.com/company/sarvam-ai)

## Other roles at Sarvam

- [Software Engineer - Model API Platform](https://feeny.ai/job/software-engineer-model-api-platform-sarvam-bengaluru-6252xtzyzy9f) — Bengaluru, India
- [Staff Engineer - Telephony](https://feeny.ai/job/staff-engineer-telephony-sarvam-bengaluru-4mt7q90ph60d) — Bengaluru, India
- [Staff Engineer - Systems Engineering](https://feeny.ai/job/staff-engineer-systems-engineering-sarvam-bengaluru-z6az3vchvm99) — Bengaluru, India
- [Growth Marketing Intern](https://feeny.ai/job/growth-marketing-intern-sarvam-bengaluru-nd6d1ycn199k) — Bengaluru, India
- [Marketing Intern](https://feeny.ai/job/marketing-intern-sarvam-bengaluru-d560v4tmah8c) — Bengaluru, India
- [Staff Software Engineer - Frontend](https://feeny.ai/job/staff-software-engineer-frontend-sarvam-bengaluru-tae371pf8h7r) — Bengaluru, India
- [GTM Strategy - Chanakya](https://feeny.ai/job/gtm-strategy-chanakya-sarvam-delhi-zp2xxvmq6s2z) — Delhi, India
- [DevOps Engineer](https://feeny.ai/job/devops-engineer-sarvam-bengaluru-5ss75qfs1rn0) — Bengaluru, India
- [Product Manager, Model APIs](https://feeny.ai/job/product-manager-model-apis-sarvam-bengaluru-jrkanrnc39wr) — Bengaluru, India
- [Visual Designer](https://feeny.ai/job/visual-designer-sarvam-bengaluru-6kcpzkwdxva6) — Bengaluru, India