Description

You will operate large-scale GPU-accelerated AI platforms, deploying and managing LLM inference systems on Kubernetes. You will own the release pipeline for production-grade LLM services, ensuring high availability and performance across multi-node clusters.

This role is on-site.

Responsibilities

• Deploy, configure, and tune LLMs using TensorRT-LLM and Triton Inference Server.
• Manage MLOps/LLMOps pipelines for automated rollouts, model versioning, and secure runtime controls.
• Implement containerized microservices and API design for scalable LLM infrastructure.
• Develop observability for inference systems, tracking GPU health, latency, throughput, and service availability.
• Apply advanced optimization methods including mixed precision, quantization, pruning, and sharding.

Required Skills

• 10+ years of experience operating containerized services at scale on Kubernetes (OpenShift).
• Deep expertise in Triton Inference Server architecture, configuration, and deployment.
• Hands-on experience with TensorRT-LLM for model optimization and inference serving.
• Proven track record of building and optimizing production-grade LLM pipelines.
• Strong background in LLMOps/MLOps processes, including engine builds and automated deployments.
• Experience with GPU-aware scheduling, load balancing, and real-time performance tuning.
• Knowledge of incident, change, and event management for mission-critical systems.

Key Skills

Education

Any Graduate