Graduate Research Intern, Machine Learning for Materials

Lila Sciences • Cambridge, MA

2d

About The Position

Join our physical sciences team to apply advanced machine learning to real problems in materials, chemistry, and physics. You’ll build models on real-world datasets generated by Lila’s laboratories, create and contribute to state-of-the-art workflows, and present crisp results that inform next steps. Example subtopics you may touch include: Bayesian optimization, representation learning, generative models, and scientific reasoning. Develop, evaluate, and ship ML models; run baselines and ablations; report clear, decision‑oriented results. Build clean, reproducible data/model pipelines (Python, PyTorch; tracking and docs). Collaborate with experimentalists, domain scientists, and engineers to turn open‑ended questions into testable milestones, working directly with real‑world data and feedback loops from Lila’s laboratories.

Requirements

• Currently enrolled in an MS or PhD program in CS, Materials Science, Chemistry, Chemical Engineering, Physics, or a related field (exceptional BS candidates with research experience considered)
• Proficiency in Python, deep learning frameworks and end-to-end workflow deployment
• Evidence of applied ML research work in physical sciences including method development, model training and evaluation, and sound code practices
• Strong communication and presentation skills, capable of conveying technical information in a clear and thorough manner
• Eager to work with highly skilled and dynamic teams in a fast-paced, entrepreneurial, and technical setting

Nice To Haves

• Experience in advanced machine learning methods related to: Bayesian optimization/active learning, representation learning/GNNs, generative models, scientific reasoning model, uncertainty & reliability, or multimodal data
• Understanding of experimental materials science techniques related to synthesis and characterization

Responsibilities

• Develop, evaluate, and ship ML models
• Run baselines and ablations
• Report clear, decision‑oriented results
• Build clean, reproducible data/model pipelines (Python, PyTorch; tracking and docs)
• Collaborate with experimentalists, domain scientists, and engineers to turn open‑ended questions into testable milestones, working directly with real‑world data and feedback loops from Lila’s laboratories.

What This Job Offers

• Career Level: Intern
• Number of Employees: 101-250 employees