ML/AI Research Internship - PhD

Description

The company builds physical AI robots for real-world industries, starting with food, mining and transport. Its systems aim to understand, predict and control physical operations with precision.

The team combines robotics, hardware, software, AI, operations, manufacturing and real estate. Robots are deployed into real environments, run there, and improved over time until they work at scale.

This internship sits on the AI/ML research side of the team and focuses on World Models and Vision-Language-Action (VLA) models for next-generation autonomous transport platforms. The work turns research ideas into systems that run on physical vehicles and help them perceive, simulate and navigate complex environments.

Responsibilities

• Research and build reinforcement learning and distillation methods for trajectory planning.
• Track new work from the wider AI community and prototype the most promising approaches.
• Build end-to-end multimodal models that take live visual input and high-level behavior goals and turn them into vehicle control.
• Build interactive world models from raw multi-sensor logs so the team can replay events and ask what a vehicle would see on a different trajectory.
• Make core autonomous driving models work reliably in new urban settings and on edge cases.
• Work with validation and research teams to run model experiments in detailed simulations, catching regressions and finding recurring performance bottlenecks.

Requirements

• Currently enrolled in a PhD program in Computer Science, EECS, Robotics, Machine Learning or a closely related quantitative field.
• Publications or submissions at top machine learning or robotics venues, such as NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, RSS, ICRA or CoRL.
• Graduate-level research in at least one of: generative world models, video generation or diffusion models; Vision-Language-Action (VLA) models or embodied AI; reinforcement learning, trajectory optimization or decision-making systems; or large-scale multimodal Transformers applied to robotics or spatio-temporal data.
• Advanced hands-on experience with PyTorch or JAX for training and fine-tuning large neural networks.
• Proficiency in Python and working knowledge of C++.
• Preferred: hands-on work with sensor fusion across multi-camera setups, LiDAR and Radar.
• Preferred: familiarity with Bird's-Eye-View (BEV) representations, 3D occupancy forecasting, multi-task learning or specialized spatial data tokenization.
• Preferred: experience with large-scale distributed training on GPU or TPU clusters, using tools such as DeepSpeed, Megatron-LM, Ray or PyTorch FSDP.
• Preferred: prior research or internship work in autonomous driving, mobile robotics or spatial AI.

Conditions

Based in the Mountain View or San Francisco office and works onsite five days a week.

Hourly pay of $70/hr.

Final pay depends on the individual's experience, skills and qualifications.

Applications are accepted on an ongoing basis.

Benefits

• The work tackles large automation challenges in the physical world, with a team that stays focused on the project from start to finish.
• The company invests in its people's trust and growth, and aims to give them a setting where they can do their best work.