Principal Developer, AI Networking

This is a fully remote position, open to applicants in California, +3 more states.

📋 Description

• Characterizing AI workloads and deep learning models designed for extensive LLM training and inference using NVIDIA supercomputers.

• This position focuses on distributed systems, particularly emphasizing high-performance networking and NVIDIA communication libraries.

• Benchmarking, profiling, and analyzing performance to uncover bottlenecks and identify opportunities for enhancements and optimizations, with a key focus on networking elements.

• Creating a PyTorch trace-based profiling, analysis, and replay toolset to facilitate benchmarking, debugging, and co-designing network systems for LLM workloads.

• Collaborating with various teams spanning hardware to software to deliver insights on performance analysis.

• Establishing performance test plans, defining performance expectations for emerging technologies and solutions, and striving to meet performance objectives.

⛳️ Requirements

• B.Sc in Computer Science, Software Engineering, or equivalent experience.

• Over 15 years of experience with high-performance networking (RDMA, MPI, NCCL, SHARP).

• Proven capability in performance evaluation techniques and methodologies.

• Familiarity with NVIDIA GPUs and the CUDA library.

• Understanding of deep learning frameworks such as TensorFlow or PyTorch.

• Proficiency in networking collective communication libraries like NCCL and protocols such as RoCE and RDMA.

• Quick and self-motivated learner with strong analytical and problem-solving abilities.

• Skilled in programming languages: Python, Bash, and C++.

• Experience in a container-based development environment.

• A strong team player who communicates effectively and collaborates well with others.

🏝️ Benefits

• Equity

• Benefits