Senior AI/ML Validation Engineer

• Own end-to-end test strategy for AI/ML workflows (PyTorch, vLLM), GPU runtimes, drivers, and firmware across kernel and user space
• Develop scalable automation frameworks spanning unit, integration, HIL (hardware-in-the-loop), system, and end-to-end tests
• Implement and maintain CI quality gates (GitHub Actions/Workflows, Jenkins), including automated build, test execution, artifact management, reporting, and flake reduction
• Design and execute performance, stress, reliability, soak, and long-haul tests targeting GPU compute, memory, I/O, and serving throughput/latency
• Validate cross-platform compatibility (Linux/Windows), covering driver interfaces, kernel interactions, firmware behavior, and runtime stability
• Create reproducible environments with containers/orchestration
• instrument telemetry and observability for data-driven QA
• Apply agentic AI techniques to accelerate test generation, triage, and root cause analysis
• integrate intelligent diagnostics into pipelines
• Develop rigorous test cases for low-level features (PCIe, DMA, interrupts, memory management), error handling, recovery, and fault injection
• Define and track quality KPIs (coverage, defect escape rate, MTTR, performance regressions) and drive continuous improvement
• Lead defect triage across hardware, firmware, driver, runtime, and model layers
• collaborate with engineering to resolve issues rapidly
• Produce comprehensive documentation: test plans, procedures, fixtures, coverage maps, readiness criteria, and retrospectives

• 8–12 years in QA/Test for systems software or platform engineering, with at least 4 years focused on GPU software, device drivers, or firmware validation
• Demonstrable ownership of validation for AI/ML pipelines and serving stacks using PyTorch and at least one modern inference framework (e.g., vLLM), including accuracy baselining and performance regression detection
• Proven expertise testing drivers and firmware with hands-on work in: PCIe fundamentals (link training, BARs, MSI/MSI-X), DMA engines, interrupt handling, and memory models
• Failure modes: error injection, recovery paths, power/thermal events, and persistence across reboot/upgrade cycles
• Deep proficiency in Linux (kernel/user space) and practical experience with Windows driver ecosystems
• ability to: Read kernel logs and symbols, trace with ftrace/perf/ETW, and perform cross-layer debugging
• Build custom kernels/modules and analyze crash dumps (kdump, WinDbg)
• Strong programming for test automation: Python for framework and orchestration (pytest or equivalent), robust mocking/fixtures, and data-driven test generation
• C/C++ for low-level test harnesses, protocol exercisers, and performance micro-benchmarks
• Bash/PowerShell for environment setup, CI scripting, and reproducibility
• CI/CD mastery with GitHub Actions/Workflows and/or Jenkins: Design gated pipelines with parallelization, artifact management, flaky test quarantine, and automated rollback criteria
• Integrate metrics, alerts, and quality reports
• enforce go/no-go release thresholds
• Performance testing rigor: Methodology for baselining, variance control, and noise isolation
• application of statistical techniques (e.g., confidence intervals, A/B comparisons) to detect regressions
• GPU-focused profiling and analysis (e.g., perf counters, memory bandwidth, kernel occupancy)
• Tooling fluency: gdb, perf, ftrace, valgrind, WinDbg, ETW
• log/trace correlation
• containerized test environments (Docker) and familiarity with Kubernetes for distributed tests
• Exploratory testing mindset: Hypothesis-driven investigation, boundary and adversarial testing, fuzzing (protocol/API), chaos/fault injection, and reverse-engineering of interfaces when documentation is limited
• Communication and leadership: Clear, concise defect reporting
• ability to drive triage across teams
• establish and evangelize quality standards
• maintain strong documentation discipline
• BS/MS in Computer Science/Computer Engineering, or related discipline

• Lab ops for QA: rack mounting, server configuration, BMC/IPMI, BIOS/fw updates, network/storage setup, power/thermal profiling
• Front-end/UI testing experience for internal tools: ReactJS, web UI automation, accessibility and usability checks
• Backend/DB validation: REST/gRPC testing, SQL/NoSQL, schema migrations, data integrity, performance tuning
• Observability: Prometheus/Grafana, OpenTelemetry
• integrating quality signals and alerts into CI/CD and release gates

AMD benefits at a glance