GSoC 2026: Scalable Online Bayesian State Space Models

Hi! I am Harshil J Soni,currently a 3rd Year Undergrad studying Computer Science.My interests lie in scientific machine learning, where mathematical modeling and computational methods come together to analyze complex systems,as it is a sweet intersection between where true discovery intersects with the ever evolving field of ML.

As I was going through the discourse over the last couple of days and was planning on contributing, although seeing the moderators highly mention that working through and getting a thorough understanding would be better, I went through a few of their tutorials. Below, along with my notes, I shared one of the case study implementations of PyMC.

Currently I am half way through this video made by @jessegrabowski , which has been a profound amount of learning, and I truly agree that jumping directly into issues would not be as helpful as this has been so far.
https://www.youtube.com/watch?v=G9VWXZdbtKQ&t=2555s

Next I plan to:

  1. Finish working through the full video and complete the accompanying examples.

  2. Go through the repository issues now that I have a better understanding of the tool.

  3. I plan to integrate the time-series analysis techniques from this approach into my project,which is one of the main reasons Bayesian State Space model particularly interest me asit can directly support my current project and ongoing research on solar storm prediction
    https://github.com/BeathovenGala/Auralis
    https://auralis-juliett.vercel.app/

    What are your thoughts or suggestions?