Director, Perception - Autonomous Vehicles

NVIDIA NVIDIA · Semiconductors · Santa Clara, CA

Director of Perception for Autonomous Vehicles at NVIDIA, leading teams to develop and deploy state-of-the-art deep learning models for real-time 3D world reconstruction and navigation. This role involves end-to-end ownership of the ML lifecycle, from data generation to deployment on NVIDIA DRIVE platforms, with a strong emphasis on safety-critical systems and cross-functional collaboration.

What you'd actually do

  1. Lead and Inspire: Manage, mentor, and scale a high-performing organization of engineering managers, applied researchers, and software engineers dedicated to autonomous vehicle perception.
  2. Drive the Architecture: Guide the strategic direction, design, and execution of modern network architectures (e.g., Transformers, BEV, Occupancy Networks, Vision-Language-Action models) for multi-sensor fusion (camera, LiDAR, radar) and real-time 3D world reconstruction.
  3. End-to-End Ownership: Oversee the entire machine learning lifecycle—from active learning, data mining, and synthetic data generation to model training, model performance acceleration, and deployment on NVIDIA DRIVE platforms (Orin/Thor).
  4. Cross-Functional Teamwork: Partner closely with leaders across Planning and Control, Mapping, Hardware, and Safety teams to ensure seamless integration of the perception stack and alignment on AV system-level metrics.
  5. Cultivate Excellence: Foster a culture of radical ownership, high-velocity execution, and continuous innovation. Ensure all algorithms and production code adhere to strict automotive quality and safety standards (e.g., ISO 26262).

Skills

Required

  • Ph.D. or master's degree in computer science, Robotics, Artificial Intelligence, visual computing, or a related field (or equivalent experience).
  • 10+ overall years of industry experience in deep learning, computer vision, or autonomous robotics, with at least 5+ years in a senior leadership role managing large teams or multiple layers of management.
  • Deep, hands-on theoretical and practical expertise with modern network architectures (Transformers, CNNs, Foundation Models) and complex multi-sensor fusion paradigms.
  • A proven track record of successfully deploying production-grade, safety-critical machine learning models to edge/embedded computing platforms.
  • Strong foundational knowledge in Python, C++, and deep learning frameworks (PyTorch, JAX, or TensorFlow).
  • Exceptional communication and leadership skills.
  • A highly motivated, entrepreneurial mindset.

Nice to have

  • Demonstrated success in shipping autonomous vehicle software to mass-production.
  • A recognized presence in the global AI/CV community, evidenced by top-tier publications (CVPR, ICCV, NeurIPS) or significant open-source contributions.
  • Hands-on experience with next-generation paradigms, such as end-to-end autonomous driving architectures, Large Vision Models (LVMs), or sim-to-real transfer techniques.
  • Deep understanding of hardware-software co-design, specifically optimizing modern networks for NVIDIA GPU architectures.

What the JD emphasized

  • production AI
  • production-grade, safety-critical machine learning models
  • shipping bleeding-edge technology
  • shipping autonomous vehicle software to mass-production
  • strict automotive quality and safety standards (e.g., ISO 26262)

Other signals

  • leading AI model evolution
  • shipping production AI
  • deploying production-grade, safety-critical machine learning models