RMG-Py is a tool for automatically generating chemical reaction networks for modeling reaction systems including pyrolysis, combustion, atmospheric science, and more. I served as the lead developer in the period from 2016 to 2018, during which I introduced two main initial releases v2.0.0 and v2.1.0.

Main items include

Parallel computing

Bicyclics decomposition of polycyclics

Reaction network pathway analysis

Deep learning based molecular property prediction

Learn More: RMG Website

DDE is a Python library for quickly training and deploying data-driven estimators for thermochemistry and kinetics. I initiated this effort in 2017. Since then, we've introduced several key features, including

Graph Convolutional Neural Networks for molecular property prediction such as entropy, heat capacity, etc.

Uncertainty estimation for deep neural networks using Dropout strategy as an approximation of Bayesian Inference.

Learn More: DDE GitHub Repo