MLJAR is a full platform for data work.

We help users go from the beginning of data analysis, through model training, to deployment.

We want to make Data Science easier, faster, and more practical.

MLJAR AutoML is an open-source AutoML framework for Python.

It helps you train machine learning models with less manual work.

MLJAR AutoML can:

• try many algorithms,
• do feature engineering,
• tune hyperparameters,
• create documentation for each model,
• explain model performance,
• check fairness and bias,
• mitigate bias,
• generate a web app from trained AutoML models.

It is not a black box. You get reports, code, metrics, and model documentation.

Mercury is an open-source framework for turning Python notebooks into web apps.

With Mercury you can:

• share notebooks with non-technical users,
• build a web app without rewriting your notebook.

You write a notebook. Mercury serves it as a web app.

Mercury apps can be deployed with Docker or with 1-click deployment on the MLJAR platform.

SuperTree is an open-source Python package for beautiful and interactive Decision Tree visualizations.

Decision Trees are loved because they are easy to explain — at least in theory. Scikit-learn is a great package, but the default Decision Tree visualization can be hard to read. SuperTree makes Decision Trees easier to explore, understand, and present.

MLJAR Studio is our main platform for working with data.

It includes an AI Data Analyst that helps you explore and analyze data in a notebook environment. The AI Data Analyst can create Python code, run analysis, make charts, and explain results in simple language.

MLJAR Studio also includes AI-assisted notebooks, so you can write, edit, explain, and improve Python code directly in your notebook.

AutoLab Experiments are AI agents for machine learning.

AutoLab works iteratively. It creates complete machine learning pipelines, runs experiments, checks scores, and looks for improvements.

It can search for:

• better models,
• better parameters,
• better feature transformations,
• better machine learning pipelines.

Each experiment is saved as a notebook, so you can inspect the code, results, and artifacts.

• Website: mljar.com
• MLJAR Studio: mljar.com/docs
• MLJAR Platform: platform.mljar.com
• MLJAR AutoML: github.com/mljar/mljar-supervised
• Mercury: github.com/mljar/mercury
• SuperTree: github.com/mljar/supertree

From data analysis to model deployment — MLJAR helps you build faster. 🚀