CUDA Software Engineer

As a CUDA Software Engineer, you will play a crucial role in developing and optimizing high-performance computing libraries and applications. Your primary focus will be on implementing cutting-edge machine learning, computer vision, and numeric algorithm solutions, ensuring their efficiency on GPUs and other hardware accelerators. Your responsibilities will include designing, implementing, and optimizing parallel algorithms using CUDA and C++ for various target platforms. Collaboration with software and hardware teams will be essential to maximize performance across different processing units. You will be expected to apply low-level optimizations such as parallelization, memory management, vectorization, and efficient data movement. Additionally, you will be responsible for building performance benchmarks, conducting profiling and analysis, and fine-tuning implementations to meet performance objectives. Understanding customer requirements and delivering tailored solutions will be a key aspect of your role. Staying abreast of the latest developments in CUDA programming, GPU architectures, and optimization techniques will also be required. To qualify for this position, you should hold a Bachelor's or Master's degree in Computer Science, Information Technology, Electronics & Communication, or a related field. You must have at least 2 years of hands-on experience in CUDA development and GPU programming. Proficiency in C/C++ programming, along with experience in CUDA, OpenCL, or similar parallel computing technologies, is essential. A solid understanding of GPU architecture, memory hierarchies, and multi-threaded programming concepts is also necessary. Having familiarity with performance profiling tools like NVIDIA Nsight and Visual Profiler will be advantageous. Experience in machine learning, computer vision, or high-performance numeric computing is preferred but not mandatory. Strong analytical, debugging, and problem-solving skills are crucial for success in this role. The ability to work independently and collaborate effectively in a fast-paced team environment is also important. Candidates with exposure to deep learning frameworks (such as TensorFlow, PyTorch), experience in optimizing algorithms for low latency and high throughput, and knowledge of SIMD instructions and cache optimization strategies will have an added advantage in this position.,