Botauto GPU Engineer at Botauto to optimize GPU performance for real-time autonomous driving workloads. Develop and optimize GPU-accelerated components using CUDA and related technologies.
Responsibilities
Optimize end-to-end GPU performance for real-time autonomous driving workloads, including sensor processing (e.g., camera, LiDAR) and neural network inference.
Develop and optimize parallel computing algorithms and GPU-accelerated components using technologies such as CUDA.
Collaborate with cross-functional teams to design and improve onboard GPU software architectures that meet the computational requirements of perception, planning, and control modules.
Profile and analyze bottlenecks across GPU computation, memory access, data movement, synchronization, and CPU–GPU interaction.
Debug and optimize GPU-based software to improve latency, throughput, resource utilization, and runtime stability on embedded platforms.
Qualification
Strong proficiency in C/C++ and PythonExperience with NVIDIA Jetson ThorExperience with model quantization
Required
Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, or a related field.
Strong knowledge of parallel computing principles, GPU architecture, memory hierarchy, and performance optimization techniques.
Experience profiling GPU applications using tools such as NVIDIA Nsight Systems, Nsight Compute, or equivalent tools.
Experience deploying or optimizing neural network inference workloads using technologies such as PyTorch, ONNX, and TensorRT.
Experience with real-time embedded systems and handling large data streams from sensors (camera, LiDAR, radar).
Strong proficiency in C/C++ and Python.
Preferred
3+ years of experience in GPU programming and optimization (e.g., CUDA, OpenCL, Vulkan).
Experience with NVIDIA Jetson Thor, NVIDIA DRIVE Thor, or similar embedded GPU platforms.
Experience with model quantization, including FP8 and NVFP4.
Experience managing concurrent GPU workloads and resource isolation using technologies such as NVIDIA Multi-Process Service (MPS), Multi-Instance GPU (MIG), or other related technologies.
Experience with GPU-accelerated sensor data compression, including camera, LiDAR, or other onboard sensor data.
