dAIve Predict · No-code machine learning for engineers

From data to prediction. Without code.

dAIve Predict takes engineers through a clear flow from data to a validated model. Surrogates for simulations, test analysis and process forecasts, in the browser or fully on premise.

Try 14 days free Request a demo

No credit card. The trial starts when you register.

Real recording of the web app · demo project NACA airfoils (lift and drag)

Models: Random Forest, XGBoost, FNN, RNN
Data: Simulation, test, process (tabular format)
Runs: Browser or desktop (local)
In use: with pilot customers since 2023

What makes dAIve Predict different

Guided, not scripted

Import, description, assessment, training and prediction follow one another. No custom ML scripting project needed.

Understand data before training

Distributions, correlations, Data Size Analysis and Input Dropout show whether the data are sufficient and which inputs matter.

Metrics, not a black box

Every model comes with a learning curve, loss and R². Cross-validation and batch training make comparisons traceable.

Separate training from use

One specialist trains the model; colleagues use it in prediction-only mode.

Local when it has to be

The desktop version works offline on your hardware. Projects and models stay with you as .dvp and .dvm files.

Built by engineers

It grew out of our own simulation projects, including injection molding and e-mobility.

The flow in four steps

Import
Load a CSV, check separator and columns, choose inputs and target.
Understand
Review distributions, correlations, Data Size Analysis and Input Dropout.
Train
Pick a model type, configure, train and check the quality metrics.
Predict
Score new cases, export results, save the model as .dvm.