Like you said these AI tools are great at little task that have been done a million times. Been using Qwen coder for writing unit test (at least a template that I can edit) and documentation. It's great for that! And it actually gets done, rather than me saying I'll get to it later 😅

Another resource to consider is github.com/rlabbe/Kalma... it helped me a lot when I first started getting into tracking/filtering. Like starting with zero knowledge of the field 😅 So depending on your current knowledge it may be less useful.

I just implement them from scratch. But Turing.jl does have a sub-project for SSMs github.com/TuringLang/S... Haven't used it yet, but it may be an option if you want a framework.

Is your data non-linear/Gaussian? If so a particle filter may be better. It's my good too method for filtering unless I have a very specific reason to use a kalman filter, e.g. limited computing resources.

Tidier is a great library. Use it and AlgebraOfGraphics.jl all the time for data analysis.