Tech Stack

Description

You will work on developing advanced generative models (e.g., trajectory transformers, diffusion models) to learn valid human driving behavior and use it as a safety ruler for autonomous vehicle releases. You'll benchmark methods and collaborate cross-functionally to establish statistical safety guarantees for large-scale simulation validation.

Requirements

• Currently PhD student in CS, Robotics, ML, Statistics or related field
• Expertise in sequence modeling with Transformers, self-attention, cross-attention for time-series/trajectory data
• Hands-on experience with generative AI paradigms (next-token prediction or diffusion models)
• Strong grasp of probabilistic ML and uncertainty quantification
• Strong Python and deep learning framework skills (PyTorch or TensorFlow)

Responsibilities

• Develop Transformer-based generative models for learning valid human driving behavior
• Establish statistical safety guarantees using conformal prediction or similar techniques
• Benchmark traditional geometric methods vs generative/ML approaches for reducing false fails
• Collaborate with cross-functional teams on multi-modal truth and probabilistic safety