2027 Summer Intern, PhD, Machine Learning - Simulation & Closed-Loop Training
Core
Design and scale reinforcement learning post-training and distillation algorithms for foundation world models in simulation to improve autonomous driving safety and realism.
Role type
PhD Summer Intern, Machine Learning (Simulation & Closed-Loop Training)
Builds
Foundation world models and closed-loop multi-agent simulators for autonomous driving
Domain
Autonomous driving, Reinforcement Learning, Generative World Models
Deliverable
production ML models
Required skills
Python, JAX/Flax, PyTorch, Reinforcement Learning (PPO, off-policy RL, constrained MDPs), Generative Modeling (Flow Matching, Diffusion Policies), Neural Architecture Design, Tensor Implementation
Preferred skills
Top-tier ML publications, Multi-agent closed-loop simulation, Sim2Real transfer, Distributed training on TPUs/GPUs, C++
Technologies
JAX, Flax, PyTorch, TPU, XLA, MuJoCo, IsaacGym, NuPlan, Waymax
Responsibilities
Design RL post-training algorithms, develop learned reward models and discriminators, run closed-loop ablation studies on TPU clusters, write modular JAX code for simulation codebases
Seniority
PhD Candidate (Summer Intern)
