Co-Op, Bayesian Methods for Complex Decision Making

Lila Sciences • San Francisco, MA

9d

About The Position

Lila Sciences is building AI systems to accelerate discovery across the physical and life sciences. Within Physical Sciences AI, this Co-Op will work alongside research scientists and engineers on a focused project exploring how AI systems can make better experimental decisions. The work supports autonomous discovery agents that select experiments for scientific campaigns involving discrete and continuous choices, multiple objectives, hard constraints, and evidence from literature and characterization data, while accounting for the cost of experiments in time and materials.

Depending on the candidate’s background, the project may involve building agents to select experiments, evaluating when in-context learners can be trusted as decision-makers, or developing Bayesian strategies for limited experimental budgets. Methods will be tested against live campaigns. The Co-Op will design and benchmark experiment-selection strategies, develop agents using surrogate models, explore Bayesian-inspired in-context learning, and conduct retrospective and live evaluations. The role includes presenting findings to machine-learning and experimental teams, with the option to contribute to a publication.

Requirements

• Pursuing a Master's or PhD in statistics, machine learning, computer science, operations research, applied mathematics, or a quantitative science.
• Solid grounding in Bayesian inference and decision-making, including priors and posterior predictive distributions.
• Research experience in Gaussian processes, Bayesian optimization, bandits, reinforcement learning, or probabilistic modeling.
• Strong Python skills, including PyTorch and probabilistic libraries such as Pyro, PyMC, or BoTorch.
• Comfort working with small, noisy, and heterogeneous datasets.
• Clear communication with machine-learning and experimental collaborators.

Nice To Haves

• Experience applying these methods to real problems in materials science, chemistry, or another physical science.
• Familiarity with in-context learning as implicit Bayesian inference, meta-learning, or prior-data fitted networks.
• Experience with optimization over categorical or combinatorial spaces, including mixed-variable, multi-objective, or batch settings.
• Experience with reinforcement learning for sequential decision-making or Bayesian experimental design.
• Experience training models on multimodal scientific data such as text, images, spectra, or simulations.

Responsibilities

• Design and benchmark experiment-selection strategies that balance exploration and exploitation under realistic budgets.
• Build autonomous discovery agents that propose the next experiment from surrogate models.
• Develop in-context learning approaches inspired by Bayesian inference and evaluate their calibration and robustness.
• Run retrospective and live evaluations on active campaign data using clear baselines and metrics.
• Present findings to machine-learning and experimental teams.
• Optionally contribute to a publication.

What This Job Offers

• Job Type: Full-time
• Career Level: Intern