Accelerate Materials Qualification with an MLOps Platform

Accelerate materials qualification with Domino's MLOps platform. Watch a scientist go from decades of institutional data to a governed qualification package for a titanium alloy, in minutes instead of months.

Search institutional memory without a command line. Domino integrates with GOOFY (Los Alamos' Grand Unified File Index) to query classified experiment data using natural language, with every result versioned and traceable before any model runs.

Screen with physics-first AI. Ten thousand synthetic compositions are filtered through first-principles metallurgical constraints before reaching any ML model. The qualification agent scores candidates, recommends processing parameters, and issues a risk verdict your program manager can fully audit.

One system of record across hybrid compute. Surrogate ML models run on Kubernetes while confirmatory simulations dispatch to an external HPC cluster via SLURM, with every job tracked, outputs auto-versioned, and results fed back into continuous retraining.

00:00 — What Is Accelerated Materials Qualification?

01:20 — What Does an AI Audit Trail Look Like Before Computation?

02:30 — Physics-First Candidate Screening: How the Filter Funnel Works 03:45 — Model Governance Framework: Qualification Agent to Multi-Lab Approval

04:45 — MLOps Monitoring Across Kubernetes and HPC (SLURM)