Product Manager, CUDA-Q Libraries
NVIDIAFull Time
Junior (1 to 2 years)
Ann Arbor, Michigan, United States
Key technologies and capabilities for this role
Common questions about this position
The role requires deep technical knowledge across CUDA, TensorRT, and embedded ML deployment, along with experience in model conversion (PyTorch/ONNX), custom CUDA kernels, C++ code, and ML inference pipelines on embedded hardware.
The team builds the embedded inference foundation for Torc’s deep learning models on NVIDIA-based platforms, develops custom CUDA kernels and pre/post-processing pipelines, and delivers high-performance inference for the autonomy stack, working closely with Perception, Application Engine, Systems, and Hardware teams.
The role involves hiring and leading a high-performance team, building a culture of ownership, collaboration, and continuous improvement, providing coaching and career development, setting technical goals and KPIs, and improving engineering processes.
This information is not specified in the job description.
This information is not specified in the job description.
Develops autonomous driving technology for trucks
Torc Robotics develops software systems for self-driving trucks, focusing on Level 4 autonomous driving technology that allows trucks to operate without human intervention in specific conditions. Their technology enhances road safety and meets the logistics industry's growing demands. Torc Robotics partners with major truck manufacturers, like Daimler Trucks, and collaborates with companies such as Luminar Technologies to integrate advanced sensors into their systems. They generate revenue by selling their software to fleet operators and truck manufacturers, while also providing ongoing support and updates. The company's goal is to improve efficiency and safety in freight transportation through their autonomous solutions.