Lila Sciences is looking to solve the challenges of sparse, noisy, and heterogeneous scientific datasets by developing robust ML models that accelerate the design and validation of novel materials, thereby advancing the mission of building an autonomous scientific superintelligence.
Requirements
- Strong proficiency in Python and modern ML frameworks (PyTorch, TensorFlow, JAX) and models in sparse, time-dependent data settings (few-shot learning, time-series prediction).
- Familiarity with materials datasets (experimental and/or computational) and performance characterization.
- Experience with time dependent data modeling methods.
- Experience with physics-informed ML or hybrid physics/ML approaches.
- Familiarity with multimodal data integration (e.g., combining simulation, imaging, spectroscopy, and tabular data).
Responsibilities
- Develop ML models to predict materials performance and reliability under diverse application conditions (e.g., stress, temperature, chemical environments, aging).
- Design data-efficient learning strategies for sparse, small, or incomplete experimental datasets.
- Integrate physics-informed priors, time-series prediction concepts, multi-modal methods and probabilistic modelling into predictive frameworks.
- Curate, preprocess, and interpret complex experimental and simulation data.
- Build scalable ML workflows that can be deployed within Lilaâs platforms.
Other
- PhD (preferred) or equivalent experience in Materials Science, Applied Physics, Machine Learning, Computer Science or related fields.
- Ability to collaborate across ML and materials science teams to deliver impactful methods and frameworks.
