GPU programming engineer

GPU programming engineer

Experience: 4+ Years

Location: Chennai, India

Employment Type: Contract role

Job Summary We are seeking an experienced GPU Programming Engineer to join our team. In this role, you will focus on developing, optimizing, and deploying GPU-accelerated solutions for high

performance machine learning workloads. The ideal candidate has strong expertise in GPU programming across one or more platforms (e.g., NVIDIA CUDA, AMD ROCm/HIP, or OpenCL) and is comfortable working at the intersection of parallel computing, performance tuning, and ML system integration.

Key Responsibilities

• Develop, optimize, and maintain GPU-accelerated components for machine learning pipelines using frameworks such as CUDA, HIP, or OpenCL

• Analyze and improve GPU kernel performance through profiling, benchmarking, and resource optimization.

• Optimize memory access, compute throughput, and kernel execution to improve overall system performance on the target GPUs.

• Port existing CPU-based implementations to GPU platforms while ensuring correctness and performance scalability.

• Work closely with system architects, software engineers, and domain experts to integrate GPU-accelerated solutions. Required Qualifications • Bachelor's or master's degree in computer science, Electrical Engineering, or a related field.

• 3+ years of hands-on experience in GPU programming using CUDA, HIP, OpenCL, or other GPU compute APIs.

• Strong understanding of GPU architecture, memory hierarchy, and parallel programming models. • Proficiency in C/C++ and hands-on experience developing on Linux-based systems. • Familiarity with profiling and tuning tools such as Nsight, rocprof, or Perfetto. Preferred Qualifications

• Familiarity with cuDNN, TensorRT, OpenCL, or other GPU computing libraries. Document # Click or tap here to enter text.

• Exposure to AI/ML workloads, computer vision, or video/image processing.

• Experience with multi-GPU programming or heterogeneous computing environments.

• Knowledge of OpenACC, SYCL, or other directive-based programming models is a plus.