MIT Short Course: Applied AI for Materials Discovery (July 27-30, 2026), LIVE ONLINE

Applied AI for Materials Discovery 

July 27–30, 2026, Live Online

Instructor: Prof. M.J. Buehler (MIT)

TL;DR: A focused, hands-on MIT short course on the future of AI-driven materials discovery, covering agentic AI, “AI Scientist” workflows, foundation models, generative AI, PyTorch/fine-tuning, RL/reasoning models and agents, diffusion and flow models, graph neural networks, LLMs/LRMs, swarm intelligence, physics-based validation, and experimental methods and scale-up. The course includes technical lectures, live clinics, hands-on labs, participant projects, and networking. Participants will earn an official MIT certificate.

This course covers the next generation of AI for science: moving beyond predictive models toward agentic, closed-loop discovery systems that can read literature and lab notes, formulate hypotheses, write and run code, connect to simulations, reason over materials data, and support physically grounded inverse design. We will cover AI Scientist architectures, planning/memory/tool use, multimodal foundation models, PyTorch and fine-tuning workflows including RL techniques, diffusion and flow-matching models for materials generation, graph neural networks, neural interatomic potentials, neural operators, LLMs/LRMs, autoresearch and loops, PRefLexOR and reflective agents, swarm intelligence (ScienceClaw/Infinite), and physics-based validation using tools such as molecular simulation and FEA.

A central theme is why this matters now: materials discovery is shifting from intuition-driven search and static ML prediction toward autonomous, verifiable workflows where AI systems can propose candidates, test them against physics and manufacturing constraints, learn from failures, and accelerate the path from concept to experimentally meaningful materials. The course also features responsible AI, interpretability, deployment, model distillation/quantization, small language models, human-in-the-loop interfaces, and lab/manufacturing scale-up.

If you need support in attending the course, a limited number of partial fellowships are available for academic applicants. Please email mbuehler [at] MIT.EDU (mbuehler[at]MIT[dot]EDU) with a copy of your CV.