Company Overview

Oracle Corporation is a global leader in enterprise software and cloud computing, powering thousands of organizations worldwide with innovative database, application, and infrastructure solutions. Our Oracle Cloud Infrastructure (OCI) division is at the forefront of delivering high-performance cloud services for AI, machine learning, and high-performance computing (HPC) workloads. Join us to shape the future of scalable, distributed systems.

Position Overview

The OCI Cluster Networking team is seeking a Software Developer 4 with deep expertise in NCCL and distributed systems to design and optimize ultra-high-performance networks and software stacks. This role focuses on scaling to hundreds of thousands of GPUs for cutting-edge AI/ML/HPC workloads, ensuring low-latency, high-throughput communication in massive clusters. You will collaborate with top engineers to push the boundaries of distributed computing performance at Oracle Cloud Infrastructure.

Key Responsibilities

• Design and implement high-performance networking software stacks optimized for NCCL and GPU-accelerated distributed training.
• Tune and optimize collective communication primitives to achieve sub-microsecond latencies across massive GPU clusters.
• Develop scalable distributed systems software that supports hundreds of thousands of GPUs for AI/ML/HPC workloads.
• Profile and debug network bottlenecks in large-scale clusters using advanced tools and methodologies.
• Collaborate with hardware vendors to integrate next-generation networking hardware like RDMA, InfiniBand, and Ethernet fabrics.
• Build fault-tolerant systems resilient to node failures, network partitions, and varying workload demands.
• Conduct performance benchmarking and modeling for cluster-scale deployments.
• Contribute to open-source projects and internal tools that enhance OCI's networking capabilities.
• Mentor junior engineers and lead code reviews for distributed systems components.
• Document designs, optimizations, and best practices for team-wide adoption.

Required Qualifications

• Bachelor's or Master's degree in Computer Science, Electrical Engineering, or related field.
• 8+ years of experience in software development with a focus on distributed systems or high-performance computing.
• Strong expertise in NCCL (NVIDIA Collective Communications Library) and GPU programming (CUDA, ROCm).
• Proven experience optimizing network performance for large-scale clusters (e.g., RDMA, GPUDirect).
• Proficiency in C/C++ and Python for systems-level programming and scripting.
• Deep understanding of Linux kernel networking, TCP/IP, and low-level protocols.
• Experience with performance analysis tools like nsight-systems, perf, and strace.
• Strong problem-solving skills and ability to debug complex, distributed issues.

Preferred Qualifications

• Experience with Oracle Cloud Infrastructure or other major cloud platforms.
• Contributions to open-source distributed systems projects (e.g., NCCL, UCX, MPI).
• Familiarity with AI/ML frameworks like PyTorch Distributed or Horovod.
• Knowledge of InfiniBand, RoCE, or Ethernet-based GPU clustering at exascale.
• PhD in distributed systems, networking, or HPC-related field.

What We Offer

• Competitive salary and performance-based bonuses.
• Comprehensive health, dental, and vision insurance plans.
• 401(k) matching and employee stock purchase program.
• Flexible time off, including unlimited PTO and paid family leave.
• Professional development stipend and access to Oracle University training.
• Modern office in Seattle with hybrid work options and relocation assistance.

About the Team

The OCI Cluster Networking team is a world-class group of engineers pioneering the next generation of AI/HPC infrastructure at unprecedented scales. We work closely with NVIDIA, AMD, and leading cloud customers to deliver production-grade solutions for trillion-parameter models and exascale simulations. Join a collaborative, innovative environment where your work directly impacts the world's most demanding workloads.