ML Ops Engineer, Chanakya

Sarvam AI

Delhi, IndiaPosted 4 months ago

Skill Required

EngineeringMachine Learning EngineerMachine LearningSOCGitHub ActionsObservabilityCD pipelinesStatisticsKubernetesPrometheusbuildingTestNGPythonDockerArgoCDdesignGitCI/CDCloudGenerative AICICDAIFulltime

Key highlights

Required experience: 3–5 years in ML engineering or MLOps with at least one production LLM or ML system in continuous operation
Core requirement: Deep expertise in model serving (vLLM, TGI, Triton Inference Server, or equivalent) and quantised model formats (GGUF, AWQ, GPTQ)
Core requirement: Experience fine-tuning and adapting models in constrained, on-prem, or air-gapped environments
Key benefit: High ownership and high impact, from day one, working on problems at the frontier of AI in India with real population-scale impact
Operational stance: Standards are uncompromising — a model failure is treated as an operational risk, not a UX problem

Role overview

The MLOps Engineer at Sarvam owns the model lifecycle across all defence and strategic sector deployments — from serving infrastructure and monitoring to evaluation pipelines and environment management. The role works across two layers: supporting Strategic Deployment Engineers in the field, and owning the model deployment infrastructure for new products being built by the product engineering team, ensuring the system is always on, always accurate, and always auditable, with uncompromising standards where a model failure is treated as an operational risk.

Responsibilities

Design and operate model serving infrastructure across on-prem and cloud deployments
Build and maintain CI/CD pipelines for model updates, rollbacks, and evaluation-gated deployments
Monitor model performance in production — latency, accuracy drift, throughput, failure modes — and build systems that surface issues before clients do
Build evaluation infrastructure: harnesses, A/B testing, and model comparison tooling for field and lab use
Manage containerised model serving in constrained, air-gapped, and edge environments
Collaborate with Data Scientists on eval pipelines; own the infrastructure layer underneath
Create runbooks and operational playbooks that Strategic Deployment Engineers can use in the field
Own incident response for model-layer failures across all active deployments

Requirements

3–5 years in ML engineering or MLOps with at least one production LLM or ML system in continuous operation
Deep expertise in model serving: vLLM, TGI, Triton Inference Server, or equivalent; experience with quantised model formats (GGUF, AWQ, GPTQ)
Experience fine-tuning and adapting models in constrained, on-prem, or air-gapped environments, including managing data pipelines and compute limitations specific to the environment
Containerisation experience with Docker, Kubernetes, or lightweight alternatives (K3s, K0s) for constrained and edge environments; familiarity with deploying across heterogeneous hardware and infrastructure configurations
Monitoring and observability using Prometheus, Grafana, or equivalent; ability to build custom eval dashboards
Python fluency; familiarity with fine-tuning workflows and model evaluation frameworks
Hands-on experience with CI/CD tooling for ML pipelines: GitHub Actions, ArgoCD, DVC, or similar

Nice to have

You've kept a production ML system running under load — and debugged it when it broke
You don't wait for things to fail; you build systems that tell you when they're about to
You write documentation that actually gets used, by people who aren't you
You treat uptime and correctness as equally non-negotiable
You understand that operational reliability is a form of trust-building
You're as comfortable optimising inference throughput as you are writing a field runbook for a deployment engineer
You take ownership of model and stack health across every active deployment — not just the ones you set up

Benefits

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

Additional details

Sarvam is building the bedrock of Sovereign AI for India, developing India's full-stack sovereign AI platform 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
The standards for the role are uncompromising — a model failure is not a UX problem, it is an operational risk
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
If you want to work on problems at the frontier of AI in India, Sarvam is the place to be