About the job
The Autonomy Subsystems team at Motional is focused on designing and evaluating modules of the autonomy software. We are seeking a talented senior engineer to contribute to evaluation of our machine learning subsystems. Our team performs a variety of evaluations of machine learning models, including offline model eval, evaluation of open and closed-loop re-simulations, and assessment of on-road performance. Our goal is to establish metrics that describe autonomy subsystem performance, evaluate against those metrics to validate performance, and provide meaningful insight to ML developers and teams, working closely alongside them to improve the performance of ML models on the car.
Responsibilities
Define, prototype, and validate advanced metrics that bring new insights to model performance evaluation for autonomy subsystems (mainly focused on perception, prediction, and planning).
Assess coverage of existing training and testing, then curate test sets that sufficiently capture Motional’s intended deployment operational design domain, across many axes (agent types, interaction types, map/spatial, etc.).
Evaluate machine learning model performance against these metrics and
Qualifications
Minimum
An engineer who has demonstrated strong cross-functional collaborative skills and is highly self-starting. This is a highly cooperative role across systems, autonomy, infrastructure, and other parts of the organization.
Proven engineer with 5+ years of experience working on high-tech safety critical systems or robotics, ideally from a systems, robotics, or similar background.
Master’s degree in relevant fields (systems, robotics, ML, CS, etc.) or Bachelor’s degree plus significant relevant in-field experience.
Skilled Python developer, familiarity with analyzing the large scale data required to train and test ML models.
Experience designing unique metrics that capture key aspects of system behavior and are used to enhance model development.
Preferred
Experience training machine learning models, familiarity with state-of-the-art ML model architectures.
Deep experience with a specific aspect of machine learning for autonomous robotics (perception, prediction, or planning).
Knowledge of C++, especially applied to safety critical systems.
Experience with systems verification and validation techniques
Experience with regression testing updates to already-deployed software in a safety-critical environment.
Experience with automotive standards (E.g: ISO 12207, ISO 26262, ISO 21448)
Experience with software quality management and assurance activities (E.g: IATF 16949, ISO 9001)
Thorough understanding of robotics systems including sensors, actuators, mechatronics and software systems
Work in cross collaboration teams to address problems and find solutions