Lead RL Architect for Autonomous Driving

NVIDIA Corporation; Santa Clara (CA); On-site; USD 224,000 - 357,000; Full time

Job summary

NVIDIA Corporation is seeking a senior specialist to build and implement advanced Reinforcement Learning algorithms for autonomous driving. You will develop scalable training pipelines and simulation environments, and collaborate with perception and planning teams to integrate RL models into the unified autonomous driving stack.

The role requires a deep RL background, experience with policy gradients, actor-critic architectures, and C++/Python for real-time systems.

Qualifications

• 12+ years of experience in the related field.
• Strong background in Reinforcement Learning, including policy gradient methods (PPO, GRPO), actor-critic architectures, on-policy and off-policy RL.
• Proficiency in PyTorch or TensorFlow and real experience with RL-related algorithms.
• Experience in C++ and Python development for real-time systems.

Responsibilities

• Build and implement brand new Reinforcement Learning (RL) algorithms for autonomous vehicle decision-making and planning.
• Develop and maintain scalable training pipelines and simulation environments for RL training.
• Collaborate with perception, and planning teams to integrate RL models into the unified autonomous driving stack.
• Benchmark RL model performance against imitation learning baselines in complex urban environments.
• Optimize and deploy RL models to production-grade automotive hardware.

Skills

Reinforcement Learning

Policy gradient methods

Actor-critic architectures

PyTorch / TensorFlow

C++ & Python

Education

BS or higher in Computer Science/Robotics/Electrical Engineering

Tools

PyTorch

TensorFlow

C++

Python

Job description

NVIDIA Corporation is seeking a senior specialist to build and implement advanced Reinforcement Learning algorithms for autonomous driving. You will develop scalable training pipelines and simulation environments, and collaborate with perception and planning teams to integrate RL models into the unified autonomous driving stack.

The role requires a deep RL background, experience with policy gradients, actor-critic architectures, and C++/Python for real-time systems.