GPU Validation,AI

Description: AI Validation Engineer embedded systems:

Key Responsibilities:
AI/ML Validation & Quality Ownership on Ubuntu & Yocto OS Variants
• Lead the Team & Exeuctie validation efforts for complex AI/ML compute stacks across multiple platforms and environments
• Define, develop, and execute comprehensive validation plans, test strategies, and methodologies for AI software and system pipelines
• Own the defect lifecycle, including reporting, triage, root cause analysis, and closure
• Ensure validation coverage across functional, performance, stress, regression, and scalability testing
End to End AI Pipeline Validation
• Validate end to end AI workflows, including:
Validate end to end ROCmLibraries
o Model training, conversion, and optimization (e.g., PyTorch ONNX)
o Inference runtimes (ONNX Runtime, TensorRT, ROCm/HIP, OpenVINO)
o Compilers and toolchains (TVM, Vitis AI, XDNA, XLA)
o Kernel execution, memory transfers, and inference accuracy
• Validate correctness, stability, and performance of AI workloads on Linux based environments (Ubuntu, Yocto)
AI Benchmarking, Profiling & Optimization
• Define and execute AI benchmarking strategies for training and inference workloads across hardware platforms
• Perform performance profiling and analysis of AI models to identify compute, memory, and latency bottlenecks
• Collaborate with compiler, runtime, and hardware teams to drive model level and system level optimizations
• Validate performance improvements across model architectures, batch sizes, precision modes (FP32, FP16, INT8), and execution paths
• Ensure performance regressions are detected early and performance KPIs are met for production releases
ML Frameworks & Compute Stack Expertise
• Validate integration, functionality, and performance of AI frameworks including PyTorch, TensorFlow, and ONNX Runtime
• Execute and validate AI/ML workloads on compute stacks such as ROCm/HIP, CUDA, OpenCL, and heterogeneous accelerators
• Analyze performance impact of framework, compiler, and runtime changes on real world AI models
Automation & Tooling
• Design and implement Python based validation and benchmarking frameworks
• Develop reusable automation for test execution, benchmarking, profiling, result analysis, and reporting
• Drive continuous improvements in validation efficiency, benchmarking scalability, and coverage
Technical Leadership & Collaboration
• Work closely with architecture, compiler, runtime, driver, and hardware teams to debug performance and functional issues
• Provide technical guidance and mentorship to junior engineers and validation teams
• Collaborate effectively with global, cross functional teams across time zones
• Communicate validation, benchmarking results, performance risks, and quality metrics clearly to stakeholders

Required Skills & Qualifications:
8 to 12 years Experience in AI Validation & Strong Python & Scripting knowledge
Technical Expertise
• Strong background in machine learning fundamentals, including:
o Deep learning
o Large Language Models (LLMs)
o Recommender systems
• Strong experience with ROCm concepts and ROCm validation
• Hands on experience with AI benchmarking, profiling, and performance optimization of AI models
• Prior experience validating AI/ML compute stacks, including:
o HIP, CUDA, OpenCL
o OpenVINO
o TensorFlow / PyTorch integrations
• Experience validating end to end AI pipelines, including model conversion, inference runtimes, compilers, kernel execution, and accuracy
• Strong Python programming skills for automation, benchmarking, and validation
• Experience profiling AI workloads to optimize throughput, latency, and resource utilization
• Strong knowledge of Linux environments, specifically Ubuntu and Yocto
Validation, Quality & Process
• Strong experience with software validation processes, defect management, and SDLC
• Experience with production level quality assurance practices and performance driven validation
• Strong focus on test quality, performance coverage, reproducibility, and release readiness
• Ability to work independently with strong ownership of deliverables
Soft Skills
• Excellent problem solving abilities with a performance driven and analytical mindset
• Strong verbal and written communication skills
• Proven ability to work effectively with global and cross functional teams

Preferred Qualifications:
• Experience benchmarking and optimizing AI workloads on heterogeneous hardware platforms (CPU, GPU, AI accelerators)
• Exposure to performance tuning of large scale models, including memory optimization and mixed precision execution
• Experience contributing to or working with open source AI and performance focused projects

Education:
Bachelors or Master’s degree in AI/ML, Robotics, Electronics, Computer Science, or a related field.