Energy & Materials Intern — Materials Representation Learning for Accelerated Design

Toyota Research Institute•Los Altos, CA

1d•$45 - $65

About The Position

At Toyota Research Institute (TRI), weâre on a mission to improve the quality of human life. Weâre developing new tools and capabilities to amplify the human experience. To lead this transformative shift in mobility, weâve built a world-class team advancing the state of the art in AI, robotics, driving, and material sciences. The Team The long-term vision of TRIâs Accelerated Materials Design and Discovery (AMDD) program is to accelerate the development of truly emissions-free mobility. Realizing this vision will require the discovery of new materials and devices for batteries, fuel cells, and more. Our aim at TRI is to merge cutting-edge computational materials modeling, experimental data, artificial intelligence, and automation to significantly accelerate materials research. Our focus is on developing tools and capabilities to enable this acceleration. We collaborate closely with a dozen universities and national labs and colleagues across global Toyota. AMDD seeks to develop and translate the newest technologies into practice, both within Toyota and the open research community more broadly. The Internship Weâre looking for a researcher/engineer who thrives at the intersection of machine learning and materials science and is motivated to develop the next generation of material representationsâintegrating signals from multiple characterizations and measured/computed properties to support accelerated materials design. In this role, youâll collaborate with TRI researchers to prototype, evaluate, and ship representation-learning approaches that can serve as a foundation for forward property prediction and inverse design workflows.

Requirements

MS or PhD (or equivalent industry experience) in computer science, applied math, materials science, chemistry, physics, or a related field.
Strong foundations in machine learning with demonstrated experience training models on real datasets.
Familiarity with modern scientific ML approaches, including representation learning, uncertainty estimation, and/or physics-informed or hybrid physicsâML modeling.
Proficiency in Python and modern ML tooling (e.g., PyTorch/JAX, experiment tracking, reproducibility best practices).
Ability to work collaboratively across disciplines and to translate research ideas into working code.

Responsibilities

Build and maintain strong baselines comparing composition-only and structure-aware representations across a set of key property prediction tasks.
Prototype and evaluate multi-view representation methods that integrate signals from multiple characterizations and properties.
Develop a reusable evaluation and reporting pipeline to assess generalization, identify failure modes, and quantify uncertainty where appropriate.
Curate datasets and implement preprocessing workflows that better capture complex materials systems and common sources of noise or ambiguity.
Contribute high-quality, well-documented code to an internal codebase and help translate results into internal reports, publications, and/or patent disclosures (as appropriate).