Role Overview
We are looking for software engineering interns to help expand our catalog of Device eXtension (Dx) APIs for math libraries. NVIDIA's accelerated computing platform has revolutionized HPC and AI. Our team develops GPU-accelerated mathematical libraries. The ecosystem is evolving towards hybrid backends, integration with Python, NumPy, JAX, MLIR, and runtime optimization. Dx APIs allow embedding highly-optimized mathematical operations. Interns will work with senior engineers on designing, developing, and optimizing math libraries for the future.
Responsibilities
- Prototype and develop numerical algorithms for high-performance math device extension libraries (cuBLASDx, cuSolverDx, cuFFTDx, cuRANDDx, nvCOMPDx, NPPDx, libmathdx) on NVIDIA GPUs.
- Analyze GPU implementation performance and identify optimization opportunities.
- Collaborate with team to understand software use cases and requirements.
- Prototype tooling to transform math libraries' APIs into LLM-friendly representations across codegen tools.
Requirements
- Pursuing MS or PhD in Computer Science, AI, Applied Mathematics, or related field.
- Strong C++ programming, debugging, performance analysis, and test building skills.
- Parallel or GPU programming experience using CUDA, OpenMP, OpenCL, SYCL or similar.
- Experience with linear algebra, FFT, or data compression algorithms.
- Collaboration, communication, and documentation habits.
Nice to Have
- Understanding of composability, fusions, compilers, programming language implementation.
- Knowledge of GPU and network hardware architecture.
- Familiarity with Python and recent CUDA developments like cuTile and CUDA Tile C++.
- Prior work in AI-assisted software development, LLMs, MCP, building RAGs or agentic tooling.
- Exposure to numerical methods, floating-point arithmetic, numerical error analysis.