2027 Summer Intern, PhD, Machine Learning - Generative World Models & Controllable Simulation
Core
Design, implement, and scale controllable generation and test-time compute algorithms for foundation world models to support autonomous driving simulation.
Role type
PhD Research Intern (Generative World Models & Controllable Simulation)
Builds
Foundation world models for Waymo's autonomous driving simulation platform
Domain
Autonomous driving, Generative AI, Simulation
Deliverable
production ML models
Required skills
Python, JAX/Flax, PyTorch, Transformers, Diffusion/Flow-Matching models, deep RL/critic networks, tensor implementation, debugging, multimodal conditioning, representation alignment, TPU clusters, agentic autoresearch workflows
Preferred skills
First-author publications at top-tier ML/CV/robotics venues, multi-agent closed-loop simulation, counterfactual/adversarial traffic generation, Sim2Real discriminator/critic stabilization, distributed training/inference profiling on TPUs
Technologies
JAX, Flax, PyTorch, TPU, XLA, Waymax, WOMD, NuPlan, MGAIL
Responsibilities
Design and implement controllable generation and test-time compute algorithms; Investigate multimodal sensor + semantic latent conditioning; Leverage learned realism critics on large-scale TPU clusters; Write clean, modular, and well-tested JAX code in core simulation codebases
Seniority
PhD Intern
