The ML metrics team focuses on data-driven metrics development for systems and subsystems to guide training and evaluation of our neural networks. The team charter is centered around two core concepts. First: how should we score the AV system as a whole in a way that is compliant with legal, ethical and ride comfort related constraints? Second: given such system level score, how do we score subsystems in a way that informs progress towards the system level goals.
What You'll Be Doing
What We're Looking For