2027 Summer Intern, PhD, Research, Multimodal Foundation Models

Waymo•Mountain View, CA

2d•Hybrid

About The Position

Waymo’s AI Foundations team develops machine learning solutions to address open problems in autonomous driving, with the goal of safely operating Waymo vehicles in many cities and under all driving conditions. The team’s areas of focus include reinforcement learning, learning from demonstration, generative modeling, Bayesian inference, hierarchical learning, and robust evaluation, and it collaborates with other research teams in Alphabet. The intern will research data and modality contributions to multimodal foundation models. Work includes running data-mixture ablations, fitting scaling laws, researching efficient data curricula for foundation-modeling tasks, and applying training data attribution and data valuation methods to quantify dataset utility and guide data curation. This is a hybrid onsite internship. Applications are accepted on a rolling basis until the role is filled.

Requirements

Currently pursuing a PhD in Computer Science, Robotics, Machine Learning, or a related quantitative field.
In-depth knowledge of state-of-the-art foundation models.
Experience with data pipelines at scale.
Strong foundation in machine learning, deep learning, and statistical modeling or optimization.
Proficiency in Python and modern deep learning frameworks such as JAX or PyTorch for training and evaluating large-scale neural networks.

Nice To Haves

Experience with Grain.
Research or practical experience with neural scaling laws, data-mixture optimization, curriculum learning, or training data attribution.

Responsibilities

Run data-mixture ablations and fit scaling laws.
Research efficient data curricula for various foundation-modeling tasks.
Research the contribution and efficiency of different modalities in a multimodal foundation model.
Explore and apply training data attribution (TDA) and data valuation methods to quantify dataset utility and guide data curation for foundation models.