Co-op, Machine Learning for Digital Twins

Your Impact at LILA

Lila Sciences builds AI systems that accelerate discovery across the physical and life sciences. Within Physical Sciences AI, our team partners with the diverse experimental groups to build digital twins of experimental campaigns, focusing on calibrated, uncertainty-aware models that enable higher-throughput, higher-quality use of Lila’s AI Science Facilities (AISF).

As an ML for Digital Twins Co-Op, you will work on building, training, and evaluating ML models for physical and experimental systems. You will get hands-on experience with operator learning, surrogate modeling, and uncertainty quantification, shipping work that directly informs how next-generation AISF experiments are designed and run.

What You’ll Be Building

• Contribute to ML models for scientific and experimental systems, focused on a well-defined digital twin sub-problem
• Build and train surrogate, operator-learning, or physics-informed models against experimental and simulation data, with mentor guidance
• Calibrate models, quantify uncertainty, and validate against data flowing from active AISF experimental campaigns
• Frame open-ended scientific questions as concrete ML tasks with clear datasets, baselines, and evaluation criteria
• Document findings and share results in cross-departmental collaboration through write-ups and presentations

What You’ll Need to Succeed

• Pursuing a Master’s or PhD in Machine Learning, Computer Science, Applied Mathematics, Physics, Materials Science, Chemical Engineering, Mechanical Engineering, Electrical Engineering, or a related quantitative field (PhD preferred)
• Strong programming skills in Python and hands-on experience with ML frameworks such as PyTorch, JAX, TensorFlow, or similar
• Experience applying machine learning to scientific, engineering, physical, or experimental systems
• Familiarity with neural operators, operator learning, spatiotemporal modeling, field prediction, dynamical systems, scientific computing, surrogate modeling, or physics-informed ML
• Ability to turn open-ended scientific questions into concrete ML tasks with clear datasets, assumptions, baselines, and evaluation criteria
• Solid foundation in model training, validation, debugging, experiment tracking, and performance evaluation
• Comfort working with messy, heterogeneous, or evolving scientific datasets
• Clear communication and interest in collaborating across ML, software engineering, and physical science teams

Bonus Points For

• Experience with modern operator-learning methods, including Fourier Neural Operators, DeepONets, graph neural operators, transformer-based neural operators, attention-based operators, physics-informed operators, or operator learning for spatiotemporal systems
• Experience with digital twins, model update, calibration, and uncertainty-aware scientific modeling, including online/offline model updating, simulator calibration, discrepancy modeling, uncertainty quantification, out-of-distribution detection, or reliability estimation
• Experience with closed-loop scientific decision-making or physical science applications, including active learning, Bayesian optimization, design of experiments, experimental decision-making, or applications in materials science, chemistry, energy systems, catalysis, batteries, electrochemistry, additive manufacturing, fluid dynamics, thermodynamics, robotics, or computational physics