ML Platform Engineer

We are looking for a specialist who can provide operational support for AI/ML compute environments. This role centers on overseeing machine-learning workloads executed on Kubernetes and Ray-based ecosystems. The position acts as the initial escalation point for issues involving training pipelines, inference services, deployment workflows, and GPU consumption. A key expectation is maintaining the stability and efficiency of the infrastructure that drives large-scale AI initiatives. The job includes participating in an on-call roster and working closely with engineering teams, researchers, and platform owners to keep the platform dependable and performant.

Key Responsibilities

• Act as the first responder for support inquiries related to the AI/ML execution environment, covering model training jobs, inference services, deployment tasks, and GPU usage.
• Provide operational coverage—including on-call shifts—for Ray and Kubernetes clusters running distributed ML jobs across both cloud and on-prem setups.
• Track system health, investigate incidents, and resolve problems linked to failed jobs, scaling issues, cluster performance degradation, or GPU contention.
• Administer GPU allocation policies and enforce usage limits across teams to maintain fair distribution and utilization efficiency.
• Support users leveraging Ray Train/Tune for distributed training and Ray Serve for scalable inference, ensuring smooth and predictable performance.
• Diagnose Kubernetes-related problems involving pods, networking, images, scheduling, or resource shortages in multi-team environments.
• Work with SREs, platform engineers, and ML practitioners to resolve infrastructure, orchestration, and dependency issues affecting ML workloads.
• Strengthen monitoring and alerting for Ray/Kubernetes systems using observability tools such as Prometheus, Grafana, and OpenTelemetry.
• Update and expand runbooks, automation utilities, and documentation to streamline incident handling and reduce repeated issues.
• Contribute to RCAs and post-incident reviews and assist in driving automation and reliability enhancements across the platform.

Required Background

• Degree in Computer Science, Engineering, or a related technical field, or equivalent professional experience.
• 5+ years in ML operations, DevOps, platform support, or similar fields handling distributed AI systems.
• Experience offering L1/L2 support and on-call coverage for Ray-based and Kubernetes-based environments running ML workloads.
• Deep understanding of Ray operations, including job orchestration, scaling behavior, and scheduling across CPU/GPU infrastructure.
• Strong practical knowledge of Kubernetes internals—control plane, data plane, RBAC, namespaces, ingress, and resource segmentation.
• Expertise in GPU scheduling and optimization, NVIDIA tooling, and related compute frameworks.
• Proficiency in Python or Go for building automation tools and debugging distributed systems.
• Familiarity with cloud-native environments on AWS, Azure, or GCP, as well as CI/CD practices.
• Hands-on experience with metrics, tracing, and alerting solutions such as Prometheus, Grafana, and OpenTelemetry, plus incident-management platforms like PagerDuty or ServiceNow.
• Understanding of ML frameworks (e.g., PyTorch, TensorFlow) and how they behave in Ray/Kubernetes-based distributed setups.
• Strong problem-solving and communication skills, with the ability to work effectively across engineering and research groups.
• Disciplined operational mindset focused on reliability, performance, and user experience.

Job Types: Full-time, Permanent

Pay: ₹3,000,000.00 - ₹4,500,000.00 per year

Benefits:

• Cell phone reimbursement
• Food provided
• Health insurance
• Provident Fund

Work Location: In person