Hummingbird (GitHub Repo)
Hummingbird can convert trained traditional ML models into tensor computations. It supports a variety of tree-based classifiers and regressors, including Decision Trees, Random Forest, LightGBM, and XGBoost. With Hummingbird, users can benefit from all the current and future optimizations implemented in neural network frameworks, native hardware acceleration, and support for both traditional and neural network models, without having to re-engineer their models.
We are the SpaceX software team, ask us anything! (Reddit Thread)
A few of the SpaceX team members who worked on the software that flew Dragon answer questions from Reddit users. SpaceX writes all of its autonomous software in C++, using only extremely high-quality open source libraries when possible and writing custom libraries when it needs to control code quality. There are many facets to error handling, with multiple levels of sensors and backups to ensure all errors are recorded and handled. The same source on Falcon is used every mission, though the software is updated fairly regularly. Configs change every mission according to variables specific to the launch. More questions and in-depth answers are available in the thread.