Senior Software Engineer, C++ and CUDA - Analytics and Data Intelligence

Overview In this role you will advance open-source GPU data-processing libraries at NVIDIA’s Analytics and Data Intelligence group. You’ll help build and optimize the core for DataFrame and database accelerators, spanning single workstations to multi-node GPU systems. Expect to drive performance, contribute to major open-source projects (cuDF, Velox, Presto) and shape scalable, high-speed data workflows. You will work with cross-functional teams to push the boundaries of parallel data processing and open-source software impact. Compensation / Benefitsequitybenefits packageremote work options ResponsibilitiesOwn development for UcxExchange in Velox (30%)Advance GPU-to-GPU communication in Presto: roadmap, performance, correctnessOptimize multi-node analytical queries with Presto GPU (40%)Lead projects in Presto, Velox, cuDF (e.g., operator completeness, GPU memory oversubscription, CPU fallback)Contribute ideas and direction for open-source data-processing libraries Key requirements8+ years of experience in Computer Science or Software EngineeringMS degree or PhD in computer science, engineering, or a related field, or equivalent experienceStrong Modern C++ programming skillsCommitment to robust, readable, high-performance codecollaboration with cross-functional teamsownership and leadership in projectsopen-source development and publication of workHigh-performance C++/CUDA programmingExperience with GPU communication and parallelismUCX and NCCL in distributed environments