NPU Diagnostics Software Engineer

Overview In this role you will design and develop diagnostics to validate AMD's NPU/AI Engine IP across pre-silicon and production stages. You’ll collaborate with hardware, firmware, driver, and validation teams to enable AI hardware through software. You will work hands-on with low-level embedded projects, automate test development, and drive debugging and qualification efforts. This position offers impact by improving reliability and performance of next-gen AI silicon in a fast-paced, cross-functional environment. Join a high-energy team that shapes how AI hardware comes to life. ResponsibilitiesDesign and develop tests to validate NPU hardware design and features in pre-silicon and post-silicon stagesDiagnose and fix hardware issues using GDB and register-level debug toolsParticipate in new APU ASIC bring-up and executionDevelop SLT and ASIC qualification suitesOptimize diagnostics software for coverage, effectiveness, and efficiencyDevelop and deploy AI-based tools and workflows to accelerate test development, failure triage, and debugBuild and debug automation tools to improve productivity Key requirementsLow-level or embedded software development with direct hardware interactionStrong C++ programming (C++11/14/17/20 a plus)Understanding of computer architecture and operating systemsLinux software development and debugging experienceExperience with debuggers (GDB, WinDbg, etc.)Familiarity with source control systems (Git, SVN, Perforce)Ability to develop and deploy AI/ML solutions to resolve engineering problemsHands-on Python and AI/ML frameworks/tools (PyTorch, TensorFlow) experienceKnowledge of ML/inference concepts and NPU/AI accelerator architecturesFamiliarity with CPU/GPU/APU architecture, parallel programming, and multi-threadingStrong debugging, testing, and communication skillsstrong communication across teamsownership and initiativecuriosity and continuous learningC++ (modern standards)Linux and embedded developmentGDB/WinDbg debugging