Are you lost in the computer world?

We help people understand and utilize their data by building outstanding Data Science tools and learning materials.

Check our GitHub

Read our articles

Share Jupyter Notebooks

Mercury is a web framework for converting Python Notebooks into interactive web applications.

Define Notebook's parameters as YAML (similar as in R Markdown)
Interactive widgets are constructed based on YAML params
Allow others to execute parametrized Notebooks
Hide Notebook's code - great for sharing with non-coders
Download executed Notebook as HTML or PDF files
Easily schedule Notebook
Send executed Notebooks in the email

Check in GitHub

MLJAR AutoML

MLJAR AutoML is a Python package for Automated Machine Learning on tabular data.

Easily try many ML algorithms: Baseline, Linear, Random Forest, Extra Trees, LightGBM, Xgboost, CatBoost, Neural Networks, and Nearest Neighbors
Build ensembles and stack models for best performance
Advanced automatic feature preprocessing and engineering
Extensive explanations and analysis for each model
Automatic documentation in Markdown and HTML formats
Works for binary/multi classification and regression tasks

Check in GitHub

React & Django

The course how to build SaaS web application from scratch with React and Django.

During the course you build a real life web application for server uptime monitoring
Machine Learning is used for anomaly detection in server response time, so you can detect server failures before they occur
User Interface is created with React and TypeScript
Backend is created with Django and Python. Background tasks are computed with Celery framework

Course Website

Deploy ML with Django

Deploy Machine Learning Models with Django web framework. It our free course that shows you how to put ML models on the internet.

Step-by-step tutorial in Python, you have access to full code to build and deploy your own ML service
Train your first ML models, use Jupyter Notebook to learn Random Forest and Extra Trees classifiers
Build prediction endpoints for multiple ML models, track models versions
Run A/B testing between new ML models and select the best performing one

Tutorial Website

Deploy Machine Learning Models with Django

Testimonials

"Mercury seems to be very interesting. Just had a try today and was able to fire up a GUI within 10 minutes that would have taken me at least 2h using Panel. Really nice thing. "

Hugo Santos Silva

"I just looked at the mljar-supervised package. It's really thorough and much more in depth than other AutoML solutions I have used in the past. The compete param is also pretty awesome! Love it. Congratulations on the great work! Not surprised it is getting rapid adoption."

Vivek Sinha

"The results of MLJAR are the most robust in this study, having the smallest gap between quantiles." (...) "MLJAR stands out with its out-of-the-box prepared reports and offers the richest XAI functionality. "

Nature Scientific Reports