Nvidia Deep Learning Engineer End-To-End Autonomous Driving 

United States, Texas 

334602850

What You’ll Be Doing:

• Design and train innovative large-scale models—including generative, imitation, and reinforcement learning—to improve the planning and reasoning capabilities of our driving systems.

• Build, pre-train, and fine-tune LLM/VLM/VLA systems for deployment in real-world autonomous driving and robotics applications.

• Explore novel data generation and collection strategies to improve diversity and quality of training datasets.

• Collaborate with cross-functional teams to deploy AI models in production environments, ensuring performance, safety, and reliability standards are met.

• Integrate machine learning models directly with vehicle firmware to deliver production-quality, safety-critical software.

What We Need to See:

• Hands-on experience building LLMs, VLMs, or VLAs from scratch or a proven track record as a top-tier coder passionate about autonomous systems.

• Deep understanding of modern deep learning architectures and optimization techniques.

• Proven record of deploying production-grade ML models for self-driving, robotics, or related fields at scale.

• Strong programming skills in Python and proficiency with major deep learning frameworks.

• Familiarity with C++ for model deployment and integration in safety-critical systems.

• Master's degree or PhD (or equivalent experience).

• 8+ years of work experience in AV or related field.

Ways to Stand Out from the Crowd:

• Experience with LLM/VLM/VLA systems deployable to autonomous vehicles or general robotics.

• Publications, open-source contributions, or competition wins related to LLM/VLM/VLA systems.

• Deep understanding of behavior and motion planning in real-world AV applications.

• Experience building and training large-scale datasets and models.

• Proven ability to optimize algorithms for real-time performance in resource-constrained environments and strong track record of taking projects from concept to production deployment.