SoC Data Path Engineer

Overview In this role you will lead post-silicon validation of AMD’s SoC data path, including fabric and interconnects across CPU, GPU, accelerator, and chiplet subsystems. You will drive validation plans, automate coverage, and diagnose issues to ensure performance, scalability, and reliability of complex multi-die products. Work closely with design, firmware, driver, and software teams to resolve hardware–software interactions and deliver robust solutions. This is a high-visibility role that supports AMD’s mission to advance computing at scale and reliability. ResponsibilitiesLead post-silicon validation and debug of SoC fabric, interconnect, and system-level data transport across graphics and compute productsPerform root-cause analysis of silicon issues involving coherency, routing, credit protocols, transaction ordering, memory interactions, and compute pipeline behaviorDevelop and execute validation plans; improve coverage via automation, tooling, fault injection, and stress-testingAnalyze scan dumps, silicon debug logs, and system telemetry to resolve hangs, stalls, synchronization failures, and protocol violationsCollaborate with design, firmware, driver, runtime, and software teams to debug complex hardware–software interactionsCoordinate cross-functional debug efforts across multiple IPs, chiplets, memory subsystems, and fabric domains to drive issue resolution Key requirementsExperience in post-silicon validation or related semiconductor roles with ownership of validation planning and milestonesStrong understanding of SoC architecture, multi-die/chiplet systems, on-chip fabrics, coherency protocols, memory hierarchies, caches, DMA, interrupts, and address translationExperience debugging silicon issues with scan dump analysis, JTAG-based tools, or similar methods; ability to root-cause hangs, stalls, protocol violations, and performance bottlenecksProficiency in C/C++, Python, Perl, or similar languages for validation automation and toolingFamiliarity with graphics and compute architectures, GPU software stacks, runtime environments, and RAS features including error detection and fault handlingBachelor in Electrical or Computer EngineeringStrong collaboration across cross-functional teamsAnalytical and methodical problem-solvingEffective communication of complex issues and fixesC/C++, Python, Perl (validation automation)JTAG-based debugging toolsScan dump analysis and silicon debug methodologies