Senior Integration Engineer, End-to-End Model - Autonomous Vehicles

NVIDIA•Us, CA

18d

About The Position

This role focuses on integrating learned driving models with various vehicle systems, including interfaces, sensor inputs, localization, mapping, and safety systems. The engineer will establish clear contracts for model inputs, outputs, timing, state-management, and runtime interfaces. A key aspect involves collaborating with model developers to enhance model quality, debuggability, runtime behavior, and deployment readiness. The position requires investigating discrepancies between model performance in development and on vehicle platforms, and optimizing model inference and surrounding software to meet strict performance requirements (latency, throughput, memory, determinism, power). The role also involves developing tools and metrics for evaluating driving quality, safety, robustness, and regression performance at scale, as well as conducting in-vehicle testing, data collection, and analysis to complete autonomous driving missions. The engineer will develop production-level code in C++ and Python, leveraging technologies like CUDA and other GPU-accelerated tools.

Requirements

PhD with 1+ year, MS with 3+ years, or BS (or equivalent experience) with 5+ years of relevant experience in Computer Science, Computer Engineering, Robotics, Machine Learning, or a related field.
Strong C++ programming, software architecture, debugging, and performance-analysis skills.
Experience with model inference technologies such as CUDA and TensorRT.
Proficiency in Python and experience working with modern machine-learning frameworks such as PyTorch.
Experience developing software on Linux and embedded or real-time operating systems such as QNX.
Experience integrating machine-learning models into complex, performance-sensitive production systems.
Ability to diagnose issues across model behavior, application software, middleware, operating systems, and hardware.
Experience with autonomous driving, robotics, ADAS, or another real-time intelligent system.

Nice To Haves

Experience deploying end-to-end driving, robotics, or embodied-AI models on production hardware.
Familiarity with model optimization, quantization, compilation, profiling, and hardware-aware neural-network design.
A track record of turning research models into robust, measurable, and maintainable product functionality.
Self-motivation, sound engineering judgment, and a passion for solving cross-functional integration challenges.

Responsibilities

Integrate learned driving models with vehicle interfaces, sensor inputs, localization, mapping, safety systems, and other autonomous driving components.
Establish clear model input, output, timing, state-management, and runtime interface contracts.
Partner with model developers to improve model quality, debuggability, runtime behavior, and readiness for deployment.
Investigate discrepancies between model behavior in development environments and on target vehicle platforms.
Optimize model inference and surrounding software to meet latency, throughput, memory, determinism, and power requirements.
Develop tools and metrics for evaluating driving quality, safety, robustness, and regression performance at scale.
Perform in-vehicle testing, collect and analyze driving data, and complete autonomous driving missions.
Develop high-quality production code in C++ and Python using CUDA and other GPU-accelerated technologies.