Agentic AI for materials discovery

Agentic AI for materials discovery is here — watch Domino's MLOps platform design battery cathode candidates from a functional specification, not a catalog search.

See a live agentic AI workflow that queries the DOE Materials Project API, generates hypothetical candidates, and ranks top results in seconds — with a traditional machine learning model called as a tool mid-agent for validated, auditable predictions.

Explore built-in model governance: every agent decision is logged with full traceability — spec inputs, model scores, and feature contributions — so your program manager can open an audit trace and see exactly why a compound was selected.

Watch model monitoring in action: Domino automatically flags prediction drift against new lab results and fires off a retraining job, keeping your AI system current as material science evolves.

00:00 — What Is Inverse Design for Battery Materials?
01:10 — How the Agentic AI Workflow Queries Real DOE Data
02:25 — Why Agentic AI + Traditional ML Beats Either Alone
03:15 — How Model Governance Logs Every Agent Decision
04:10 — Forward Design Mode: Adjusting Composition in Real Time
04:54 — What Model Monitoring Looks Like in a Mission-Critical MLOps Platform