This application allows to specify flexible rules to generate random data with certain parameters.
Along with simpler generators (like normally distributed data), there is a special "script" generator -
the user can provide his own script in C#, that would generate data by custom rules.
DynamicCodeExecutorSH.cs - compiles such a script at runtime.
StringScriptGenerator.cs - whole routines to use a custom script.
DPPanel.cs - a base class for UI elements with data generators. Designed in a generic way to enable extending the application with arbitrary generators.

Builds a live 3D-sphere of arbitrary images (passed as arguments).
ImageSphere_render.png shows the result.
It was a part of a web application with rendering on GPU.
The code makes use of the spherical coordinate system to position images correctly.

This solves a mathematical problem from a HeadHunter challenge.

Several modules from an iOS app that was launched on the AppStore.
SHDataHandler - routines for getting and storing data into DB and related logic.
SHMainMenuViewController - view controller for the main app screen.

A module for splitting long system messages into parts.

A piece from an assignment in Matlab for Andrew Ng's ML course.

A dancing Pololu 3pi robot, based on Arduino.
Dance is performed using two wheels, for a "Romeo and Juliette" song by Dire Straits.

Solution for the "Predict Future Sales" Competition on Kaggle:
https://www.kaggle.com/c/competitive-data-science-predict-future-sales
pfs_submit.ipynb is a full solution in a Jupyther Notebook.
pfs_xgb.py is part of an extended solution in pure Python. (xgb stands for XGBoost).