Senior Software Engineer — Autonomous Driving
NVIDIA, Germany, Remote
The role
We are looking for a highly motivated Machine Learning Engineer based in Stuttgart, Germany to join our autonomous driving team. This role focuses on both classical prediction, decision-making, planning, and control components, as well as cutting-edge end-to-end solutions powered by Large Language Models (LLMs). You’ll collaborate with teams to deliver robust, high-performance software solutions tailored to the competitive demands of the different markets in the world.
What you will do
- Collaborate with engineers across NDAS teams to perform simulation and in-vehicle testing and troubleshoot issues in end-to-end Large Language Model (E2E LLM) and classical Planning & Control (PnC) components.
- Optimize classical decision-making, planning algorithms and advanced E2E LLMs for NDAS functions in complex traffic scenarios.
- Integrate, adapt, and fine-tune NDAS solutions across various OEM platforms to enhance competitiveness and reliability.
- Take function-owner responsibility for selected advanced driving functions, coordinating algorithm development, OEM integration, and functional validation from initial implementation through production readiness.
- Drive system-level issue resolution across prediction, decision-making, planning, control, and E2E ML components: use logs, recorded-data replay, simulation, and vehicle testing to isolate root causes and coordinate fixes with customer and NVIDIA engineering teams.
- Perform hands-on in-vehicle testing to evaluate driving behavior and measure functional performance in complex traffic scenarios, translating test findings and corner-case analysis into prioritized algorithm improvements and actionable feedback for model-training teams.
- Stay current with the latest academic research in E2E LLM domain in autonomous driving and quickly apply new findings to practical implementations.
What you need
- Master’s or Ph.D. in Computer Science, Robotics, Automotive Engineering, Electrical Engineering, Mechanical Engineering, or a related field.
- 8+ years of hands-on experience with at least 3 years of end-to-end learning models, especially vision-language models (VLMs) and large language models (LLMs).
- Solid understanding of trajectory generation, motion control, and decision-making in complex traffic environments.
- Proficiency in classical PnC algorithms such as Dijkstra, A*, RRT, DDP, MCTS, and ILQR.
- Strong programming and debugging skills, with expertise in C++ and Python, and a solid foundation in data structures and advanced algorithms.
- Willingness to travel for customer collaboration, vehicle integration, and in-vehicle testing.