Intellegens

Founded 2017

Design experimental workflows and predict material properties using sparse research datasets

You can use Intellegens to model complex experimental data and design laboratory experiments for materials, chemicals, and manufacturing processes. The software trains machine learning models on sparse or noisy datasets where traditional methods struggle. You can predict output properties from input ingredients, test virtual formulations to reduce physical testing cycles, and identify critical variables driving chemical performance. Research teams can guide catalyst formulations, refine alloy manufacturing parameters, and quantify prediction uncertainties to make reliable data-backed development decisions across diverse industrial research projects.

What Intellegens does

Predict output properties for unmade chemical formulationsDesign targeted experiments to collect informative data pointsCalculate uncertainty scores for model predictionsIdentify key ingredient interactions driving product failureSimulate virtual formulation trials on existing research databasesSelect process parameters for additive manufacturing machineryBuilds predictive models directly from sparse and noisy experimental datasetsQuantifies model uncertainty using nonparametric probability distributionsGenerates adaptive designs of experiments to decrease lab testing workloadsRuns virtual experiment simulations to test formulation ideas before physical mixingIdentifies which input parameters most strongly influence target chemical propertiesSuggests ingredient combinations and processing parameters to reach target specifications