PythonCUDAKubernetesPyTorchDistributed SystemsGPU Clusters

Role Briefing: Join NVIDIA's Accelerated Computing and Deep Learning infrastructure group in Bengaluru to build hyper-scale distributed training orchestration platforms for next-generation generative AI foundation models.

About NVIDIA Accelerated Computing Bureau

NVIDIA is pioneering the era of accelerated computing and generative artificial intelligence. The AI Infrastructure engineering group designs the software fabric that coordinates tens of thousands of Blackwell and Hopper GPUs across high-performance sovereign clusters.

Key Responsibilities

Architect, build, and optimize distributed orchestration pipelines for trillion-parameter LLM pre-training and high-throughput inference engines.

Collaborate with AI research scientists to diagnose network bottlenecks, GPU memory contention, and tensor parallelism efficiency using NCCL and CUDA-X libraries.

Automate cluster provisioning, telemetry monitoring, and fault-tolerant failovers across bare-metal Kubernetes and Slurm environments.

Maintain rigorous performance benchmarks for microservices handling billions of real-time inferencing tokens per hour.

Candidate Qualifications & Skills

Bachelor's or Master's degree in Computer Science, Electrical Engineering, or equivalent practical experience.

5+ years of hands-on systems programming experience in Python, C++, or Go.

Deep familiarity with Linux kernel optimization, high-speed InfiniBand/RoCE networking, and container orchestration (Kubernetes).

Proven track record with distributed deep learning frameworks (PyTorch, DeepSpeed, Megatron-LM).

Compensation & Benefits

Competitive base compensation with annual stock grant (RSUs) and performance bonuses.

Comprehensive family health insurance, wellness stipend, and relocation assistance to Bengaluru.

Flexible hybrid work schedule and dedicated GPU compute credits for personal research projects.

← Return to All Tech Positions