ML Ops Engineer, Chanakya

Ovii's Interpretation of the Role

Sarvam seeks an MLOps Engineer to own the full model lifecycle for strategic deployments, building and operating serving infrastructure, CI/CD pipelines, monitoring, and incident response. The role works on‑prem and cloud environments, collaborating closely with data scientists and deployment engineers.

Role Snapshot

• Own model lifecycle across deployments
• Design & operate serving infrastructure
• Build CI/CD pipelines for model updates
• Monitor latency, accuracy drift, and throughput
• Create evaluation harnesses and A/B testing tools
• Manage containerised serving in edge/air‑gapped settings
• Write runbooks and lead incident response
• Partner with data scientists on eval pipelines

Must-Have Requirements

• Model serving platforms (vLLM, Triton, TGI)
• Containerisation (Docker, Kubernetes)
• Monitoring/observability (Prometheus, Grafana)
• Python programming
• CI/CD tooling for ML (GitHub Actions, ArgoCD, DVC)
• ML engineering
• MLOps
• production LLM systems

Nice-to-Have Signals

• Quantized model formats (GGUF, AWQ, GPTQ)
• Edge & air‑gapped deployment experience
• Evaluation infrastructure (A/B testing, harnesses)
• Lightweight k8s alternatives (K3s, K0s)
• quantized model formats
• edge/air‑gapped environments

Work Setup

• Location: Delhi, India
• Work mode: ONSITE
• Employment type: Full-Time

Not Specified in JD

• Visa sponsorship
• Salary range
• Remote eligibility
• Education requirement
• Certifications
• Relocation
• Notice period
• Travel
• Security clearance
• Coding test
• Portfolio
• GitHub
• Writing sample
• Cover letter

What You'll Likely Work On

• Design and operate model serving infrastructure for on‑prem and cloud environments
• Build and maintain CI/CD pipelines that support model updates, rollbacks, and evaluation‑gated releases
• Monitor production model performance—latency, accuracy drift, throughput—and build proactive alerting
• Develop evaluation harnesses, A/B testing frameworks, and model comparison tooling
• Manage containerised model serving in constrained, air‑gapped, and edge environments
• Author runbooks and operational playbooks for field deployment engineers
• Own incident response for model‑layer failures across all active deployments
• Collaborate with data scientists to own the underlying evaluation infrastructure

Good Fit If You Have

• Hands‑on experience keeping a production LLM system running under load
• Proactive mindset: builds systems that surface issues before they impact clients
• Strong documentation habits that are used by non‑author teammates