Description

Key Skills: GPU, CUDA, AI/ML, PyTorch, Software Validation, Performance Benchmarking, Kernel Debugging, End-to-End Testing, Accelerator Architectures

Roles & Responsibilities:

• Validate the AI software stack end-to-end by developing and executing representative ML workloads as a proxy for real-world users.
• Write, port, and optimize AI/ML workloads to identify performance bottlenecks, functional gaps, and usability issues.
• Develop and maintain a comprehensive validation test suite covering model serving, kernel compilation, runtime, and framework integration layers.
• Reproduce, triage, and analyze issues across the stack -- from Python-level framework behavior to low-level compiled kernel correctness and GPU performance.
• Benchmark and profile workloads to monitor regressions and validate optimization improvements across releases.
• Collaborate with compiler, runtime, hardware, and framework engineering teams to provide actionable feedback and drive issue resolution.
• Design and implement scalable test frameworks and automation infrastructure to enable continuous integration and validation.
• Ensure robustness, performance, and reliability of AI workloads on custom hardware and accelerator platforms.

Experience Required:

• 5-10 years of experience in AI/ML systems, GPU programming, or software validation.
• Strong hands-on experience with CUDA, GPU architectures, and ML frameworks such as PyTorch.
• Experience profiling and optimizing ML workloads for performance and latency.
• Solid understanding of AI software stacks including compilers, runtimes, and framework integration.
• Exposure to hardware accelerators and performance benchmarking tools is an added advantage.
• Strong debugging skills across multi-layered software systems.

Education: B.Tech M.Tech (Dual), B.Tech

Key Skills

Education

Any Graduate