Motional is looking to make self-driving vehicles a safe, reliable, and transformative reality by building the intelligence responsible for understanding complex driving environments and generating safe and comfortable trajectories for their robotaxi fleet.
Requirements
- In-depth understanding of common Machine Learning and Deep Learning algorithms, especially transformer-based architectures, autoregressive decoder architectures etc.
- Experience designing, training, and analyzing neural networks/transformer-based models for at least one of the following applications: motion planning, object detection, sensor fusion, motion prediction, and/or multi-object tracking, multi-agent predictions.
- Knowledge of software engineering principles, including software design, source control management, build processes, code reviews, and testing methods
- Knowledge or course experience of generative models such as VLM, VLA etc.
- Experience with PyTorch or other Python-based deep learning frameworks
- Experience working with large data sets, data curation, and augmentation.
- Strong programming skills in C++ and/or CUDA programming
Responsibilities
- Designing and training large models for predicting surrounding agents, AV behaviors, and end-to-end planning systems by executing high-value experiments based on collaborative input from other machine learning engineers
- Prototyping and implementing continuously improving metrics to evaluate the performance of our behavioral models across many desired scenarios and robotaxi capabilities
- Timely release of major model updates, including analysis of both off-line and on-road evaluation data
- Staying up to date with the latest trends in our fast-moving industry, and proposing new architectures and network designs
- Maintaining a high-quality training and evaluation code-base, including portions of our training stack dedicated to dataset generation and evaluation
Other
- Masters or PhD in Machine Learning, Generative AI, Computer Science, Applied Mathematics, Statistics.
- This opportunity can support remote work within the United States, with occasional travel.
- Publications in relevant conferences (CVPR, ICML, NeurIPS, ICCV, ECCV, ICRA, IROS etc.)
- Project or internship experience on VLM, VLA and other foundation models for robotics and self-driving.
- Project experience working on autonomous vehicles or related fields
