Shape the future with us!
Your Role & Responsibilities
- Own the development of a high-performance perception platform powering humanoid robotic systems.
- Design and implement high-throughput, low-latency sensor pipelines in modern C++.
- Optimize GPU data paths using CUDA, including memory management, concurrency, and zero-copy techniques.
- Deploy and optimize perception workloads on embedded GPU platforms (e.g., NVIDIA Jetson).
- Translate state-of-the-art perception research into robust, production-ready systems.
- Integrate accelerated inference runtimes (e.g., TensorRT).
- Design clean and stable interfaces to downstream autonomy modules.
- Build pipelines for automated dataset generation, model training, and evaluation in collaboration with MLOps.
- Establish strong engineering practices, including automated CI testing, observability, and performance profiling.
Required Technical & Professional Expertise
- MSc or PhD in Computer Science or a related field.
- Expert-level C++ skills with strong systems fundamentals and at least 5 years of relevant experience.
- Experience building real-time, production-grade data pipelines.
- Strong understanding of GPU-accelerated systems, including efficient data movement, concurrency, and performance optimization with CUDA.
- Experience deploying software on embedded GPU platforms (e.g., NVIDIA Jetson).
- Experience with automated testing, validation, and benchmarking.
- Proven track record of bringing perception systems from research into reliable production deployments.
- Experience with frameworks such as ROS2, GStreamer, DeepStream, NVIDIA GXF/NITROS, or similar.
- Familiarity with synthetic data generation and simulation platforms (e.g., NVIDIA Isaac Sim) is a plus.
- Strong engineering judgment with the ability to balance deep technical work and pragmatic execution for high-impact results.