- Experience
- 3 – 5 years
- Pay
- Not disclosed
- Education
- Not specified
- Location
- Delhi, India
- Work mode
- On-site
- Job type
- Full-time
About this role
Sarvam is building India’s full-stack sovereign AI platform. The MLOps Engineer owns model serving, monitoring, evaluation and deployment infrastructure for defence and strategic-sector deployments, including air-gapped environments.
What you will do
- Design and operate model serving infrastructure on-prem and in cloud
- Build CI/CD for model updates, rollbacks and evaluation-gated deploys
- Monitor production model performance (latency, drift, throughput)
- Build evaluation harnesses, A/B testing and comparison tooling
- Manage containerised serving in constrained and edge environments
- Create runbooks and own incident response for model-layer failures
What the employer expects
- 3–5 years in ML engineering or MLOps with at least one production LLM/ML system
- Deep expertise in model serving (vLLM, TGI, Triton) and quantised formats
- Experience fine-tuning and adapting models in constrained/on-prem settings
- Containerisation with Docker/Kubernetes (or K3s) for edge environments
- Monitoring with Prometheus/Grafana and custom eval dashboards
- Python fluency and CI/CD for ML pipelines
Skills
- vLLM
- TGI
- Triton
- Docker
- Kubernetes
- Python
- Prometheus
- Grafana
- CI/CD
- GGUF
- AWQ
- GPTQ
Apply on the employer's site
The application opens on Ashby. Dev Portal Notes does not collect your application or charge a fee.
Know someone who fits this role?
Applications are handled on the employer's own site.
Prepare for this role
Close the skill gap
Udemy
The Python Mega Course
Python taught by building real applications, from scripting through web apps and data work.
Get the coursePluralsight
Library of tech courses
Explore expert-led courses on programming, DevOps, AI and more with Pluralsight's monthly plan.
Browse coursesSome links on this page may generate a small commission at no extra cost to you.