About the job
We are looking for a passionate and energized engineer to accelerate the integration of APIs across the CUDA-X math library software stack into modern LLM and agentic AI workflows. Leading organizations globally are using GPU-powered data centers for AI, data analytics, and scientific simulations, driving advancements in fields like LLMs, computer vision, CAE, EDA, and autonomous vehicles. Our team develops the GPU-accelerated libraries and SDKs essential for these technologies.
Responsibilities
Drive the architecture and implementation of math libraries APIs and documentation structures designed to be LLM-first: introspectable, explainable, and easily synthesized by AI agents.\\\nWork with internal and external stakeholders to deliver timely LLM-enhanced library releases.\\\nBuild metrics, tools and processes to measure impact and quality of LLM/agentic code generation.\\\nPrototype tooling and solutions to help transform math libraries’ APIs into LLM-friendly representations across popular codegen tools.
Qualifications
Minimum
PhD or MSc’s degree in Computational Science, Computer Science, Applied Math, or related science or engineering field of study is preferred (or equivalent experience)\\\n3+ years of experience\\\nRobust knowledge of LLMs, finetuning, RL, building RAGs, MCP, and building agenting tooling\\\nProven experience designing clear, composable APIs and writing high-quality, well-documented code for complex technical domains\\\nAbility to prioritize multiple projects and work independently with minimal direction\\\nExcellent collaboration, communication, and documentation habits
Preferred
Prior work in AI-assisted software engineering, code generation, or programming language design\\\nFamiliarity with CUDA-X math library APIs: cuBLAS, cuFFT, cuSOLVER, cuSPARSE, etc.