The Driver Refinement and Validation (DRV) organization’s goal is to improve the Waymo Driver performance by owning the end-to-end driving behavior for various onroad problems clusters. We work with upstream and the downstream teams on developing the ML models and integrating into the Waymo Driver. As part of the data flywheel team you will take advantage of millions of miles of driving data from a diverse set of sensors to (1) develop methods for efficiently and continuously learning from large scale real-world data, to (2) develop new post training methodologies (3) explore state of the art modeling techniques (4) build tooling and infra to democratize the use of these technologies.
You will:
Design, develop, and implement ML solutions for autonomous vehicle planning including, but not limited to, modeling, reward/loss shaping, finetuning, reinforcement learning, data shaping, etc.
Develop model evaluation recipes and metrics for measuring and improving performance of pre-trained and fine-tuned models
Develop data mining techniques to curate dataset samples from large corpus of driving data
Develop novel ways to learn from unsupervised driving data. Be able to apply SOTA techniques within Waymo’s tooling/infrastructure.
Leverage large language models and world foundation models for reasoning, decision making and multi-modality generation.
Stay up-to-date with the latest advancements in autonomous driving and machine learning, and be able quickly prototype, experiment and deploy novel SOTA solutions.
Build data tooling/infrastruction to unblock experimentation.
Collaborate closely with partner teams such as perception, research, simulation, labeling, and evaluation.
Qualifications
Bachelors in Computer Science or a similar discipline, or an equivalent amount of deep learning experience
5+ years experience in Machine Learning and/or Computer Vision
Proficiency in Python and at least one deep learning frameworks (e.g. PyTorch, JAX, or TensorFlow)
Experience in applied machine learning including deep learning, feature engineering, loss/reward shaping, data shaping, fine tuning and model evaluation
Experience with post-training techniques including such as reinforcement learning, close loop training, RLHF, etc.
Experience in dealing with large scale models and datasets
Experience in integrating ML models into production systems.
MS or PhD Degree in Machine Learning, Robotics, Computer Science or a similar discipline
Publications at top-tier conferences like CVPR, ICCV, ECCV, ICLR, ICML, ICRA, IROS, RSS, NeurIPS, AAAI, IJCV, PAMI
Experience in the autonomous driving domain, including areas like motion planning or perception
Experience with Large Language Models (LLM) or Vision Language Models (VLM), prompt engineering and chain of thought reasoning