Principal Perception Engineer, Obstacle Foundation Models - Autonomous Vehicles
Core
Lead the design and productization of NVIDIA's next-generation autonomous driving perception stack, specifically focusing on 3D obstacle perception using deep learning.
Role type
Principal, hands-on IC with broad technical leadership
Builds
3D obstacle perception models and systems for autonomous vehicles
Domain
Autonomous Vehicles / Computer Vision / Deep Learning
Deliverable
production ML models
Required skills
Deep learning architecture design, 3D perception (BEV, multi-view geometry), multi-modal sensor fusion (camera, radar, lidar), transformer-based models, large-scale pretraining, parameter-efficient fine-tuning (LoRA), CUDA development, Python, C++, PyTorch
Preferred skills
Experience deploying DNN pipelines on embedded/real-time platforms, vision-language models (VLMs), strong publication record in CV/autonomous systems
Technologies
PyTorch, CUDA, CNNs, Transformers, BEV, LoRA
Responsibilities
Own technical vision and roadmap for 3D obstacle perception; Design and develop advanced 3D perception models using multi-camera/multi-sensor inputs; Lead development of efficient production-grade deep learning models; Define KPI frameworks and analyze datasets to improve accuracy/robustness; Lead data strategy for perception including labeling and active learning; Partner with safety and systems teams for deployment readiness; Provide technical leadership and mentorship to engineers
Seniority
Principal, hands-on IC
