Software stacks

ROS2

Memory management in ROS. instead of having many pods running small parts of ROS, so many instances taking up the same memory footprints. possibly design it in a modular fashion where we say this node in cybernetics will run an ROS node with a certain configuration and connection to different system devices.

Numpy [Numba]https://numba.readthedocs.io/en/stable/index.html() ( for CUDA/GPU acceleration in python )