NVIDIA
Senior Radar Perception Engineer, Obstacle Foundation Models - Autonomous Vehicles
$224K–$356K
US, CA, Santa Clara +2 · Senior · Full-time · Posted Oct 2, 2026
- Location
- US, CA, Santa Clara
US, CA, Remote
US, Remote - Experience
- 12+ years.
- Education
- BS/MS/PhD in Computer Science, Electrical Engineering, Robotics, or related fields (or equivalent experience).
About the job
Design, develop, and productize radar-based 3D obstacle perception and multi-sensor fusion for autonomous driving, including foundation models and production deep-learning systems.
What you'll do
- Develop radar perception architecture and models for obstacle detection, tracking, and BEV scene understanding, including radar-camera-lidar fusion.
- Evaluate radar sensors and improve data strategy, KPIs, and model performance through experiments and error analysis.
- Integrate efficient, production-grade perception software with safety, systems, and software teams for deployment.
What they ask for
- 12+ years of hands-on experience developing deep-learning perception, radar signal processing, or closely related systems.
- Proficiency with frameworks such as PyTorch and experience taking models from prototype to production.
- Strong programming skills in Python and/or C++, with experience building reliable, high-performance production software.
- Proven data-driven development experience, including collaboration on radar data strategy, labeling quality, and iterative model improvement.
Nice to have
- Experience deploying radar or multi-modal perception for autonomous driving or robotics at scale; embedded or real-time DNN optimization.
- Radar physics and digital signal processing expertise; research publications or contributions; CUDA or GPU-acceleration experience.
Pay
224,000 USD - 356,500 USD (base salary)
Tools
PyTorch, Python, C++, Transformers, LoRA, CUDA, FMCW, CFAR
AI focus
Develop radar-centric foundation models and deep-learning perception, using techniques such as large-scale pretraining, weakly supervised learning, cross-modal distillation, and model-assisted auto-labeling.
Model-written summary · the original posting is authoritative