Principal Software Engineer - AI Infrastructure, OCI: Oracle

This position is for a senior-level individual contributor who will design, deploy, and operate a large-scale global Oracle Cloud Infrastructure (OCI). The role is highly technical and multi-disciplinary, requiring a deep understanding of networking, programming, and performance analysis. You'll be working on groundbreaking solutions from the ground up to ensure the high performance of AI infrastructure.


Responsibilities
Develop tools for performance benchmarking.

Conduct systematic performance studies on RDMA-backed AI GPU clusters, focusing on network performance.

Troubleshoot and resolve performance problems on RDMA clusters.

Validate the network performance of RDMA clusters in various contexts, such as with NCCL.

Document new tools and procedures, and write reports to disseminate findings.

Mentor junior engineers and contribute to network solution design and roadmap development.

Required Skills & Qualifications
Experience: 6 to 10+ years of relevant experience, with a BS or MS degree in Computer Science or a related field.

Networking: Experience with RDMA Networking (RoCE or InfiniBand).

Distributed Systems: Experience working on large-scale distributed systems, with a preference for benchmarking and optimizing parallel workloads on clusters.

AI/HPC Stack: Experience with typical elements of the AI/HPC software stack, including job schedulers, parallel file systems, and ML frameworks.

Programming: Expertise in a scripting or compiled language, with Python being preferred.

Troubleshooting: Experience with performance troubleshooting on clusters.

Cloud: Experience architecting or developing solutions on a public cloud platform.

Compensation & Benefits
Salary Range: The hiring range is from $96,800 to $223,400 per annum.

Additional Compensation: The role is eligible for bonus and equity.

Benefits: Oracle offers a comprehensive package, including medical, dental, and vision insurance, a 401(k) with a company match, paid time off, paid parental leave, and an Employee Stock Purchase Plan.