Developer - I/O Acceleration

What you'll do

• Design and build accelerated I/O paths for data-intensive analytics workloads.
• Optimize end-to-end throughput by eliminating copies, syscalls, and stalls in the storage to GPU pipeline.
• Integrate with GPU-aware runtimes and high-bandwidth fabrics like GPUDirect Storage and RDMA.
• Develop benchmarks to identify performance bottlenecks and improve I/O efficiency.
• Instrument data paths to measure cost-per-query, bandwidth-per-GPU, and CPU overhead metrics.
• Collaborate on API design for accelerated I/O usability with query engine and storage teams.

What we're looking for

• Strong modern C++ skills and deep understanding of Linux systems internals.
• Experience in storage I/O subsystems, decompression, or query-engine data paths.
• Working knowledge of GPU-aware pipelines and acceleration frameworks like CUDA.
• Proficient in performance profiling and bottleneck isolation techniques.
• Familiarity with distributed data systems and large-scale production environments.
• Production experience with GPUDirect Storage, RDMA, or NVMe-oF integrations.
• Track record of delivering software in Agile environments and contributing to CI/CD pipelines.