Advancing material manufacturing with data, ML & AI

Material manufacturing,
advanced by data.

Leucite provides tailored, domain-embedded models to accelerate root cause analysis, tighten process control, achieve target material properties faster, and more.

Material Intelligence, Applied
Why Leucite

Deep material expertise, built into every model

Deep material expertise and manufacturing know-how

ML must be grounded in the physical reality of material manufacturing. No off-the-shelf AI.

White-glove implementation

Processes, legacy systems, and teams are unique to you. We tailor our offer to your needs.

Case Study

Solving real problems with real material manufacturing constraints

Built on a lean dataset: approximately 80 datasets across 20 production stations, more than 20 input parameters, and 5 output parameters

Leucite revealed key trends in manufacturing data, linked process parameters to specific material properties, discovered the process controls most crucial to success, and suggested process optimization.

~80
Datasets
20
Production Stations
>20
Input Parameters
5
Output Parameters
Capabilities & Impact

What Leucite delivers

  • Instant parameter ranking
  • Dashboard accessible data and visualization
  • Predict and prevent potential failures
  • Tune material properties
  • Increase yield and eliminate defects
92%
Accuracy
Predicting production success
200h+
Time Savings
Root-cause analysis engineering
Process Flow

From pilot to large scale deployment, Leucite makes data and ML work for your business

01

Problem Definition

Examples: defect reduction, yield optimization, process control, material R&D, process scale-up.

02

Data Capture

Relevant data only, across systems and sites including Excel, Access, enterprise software, SQL, historians, etc.

03

Data Processing

Critical cleaning and curation, strongly guided by industry experience. Data engineering as needed for ML.

04

Analysis & Visualization

Identifying trends, critical variables, linear correlations. Determining physical anchors for ML models.

05

ML Model Training

Library of ML algorithms and relationships. Identifying critical parameters & verifying model output.

06

Act & Iterate

Share model insights through dashboard. Achieve proactive process control and improved accuracy with iterations.

Our Team

20+ years of material expertise, built into every model

Katie’s headshot goes here

Katie Colbaugh, PhD

Founder / CEO

Crystal growth expert with over a decade of industrial and academic research experience in crystal production, process engineering, analysis, manufacturing, coding, automation, and new product development.

Let’s talk about your process

Share what you’re building and what you’re working with, we’ll take it from there.

News & Events →