Hello everyone — I’m Hridayesh Chettri from Darjeeling, India. I graduated in Physics with minors in Mathematics and Computer Applications. Presently, I’m keenly interested in Bayesian time-series and probabilistic programming as I am trying to delve into Machine Learning and Artificial Intelligence.
I have had experience with mathematical computation using python in my graduate studies. For my final-year project, “Risk Management in Stock Markets and Portfolio Optimization Using Physics-Inspired Machine Learning”, I used Python to build physics-inspired models for portfolio analysis. This work convinced me that modelling uncertainty via Bayesian approaches is essential for good risk estimates and decision making.
I’ve installed PyMC, run some of the state-space examples, and watched “Getting Started with PyMC” by Chris Fonnesbeck . I am plannig to watch and study the “Statistical Rethinking” resources recommended by the mentor to the past contributors. While I am new to open-source contributions, I am actively learning the process (forking, branching, pull requests, documentation contributions) and preparing my first small PR. GSoC is an opportunity to not only contribute meaningfully to PyMC, but also to grow under mentorship and put forward good research.
I am especially interested in the project “Scalable Online Bayesian State Space Models” because it matches well with my background in physics-based modelling and time-series analysis. I find the idea of updating models step-by-step(dynamic modelling) as new data arrives very interesting and very essential in handling uncertainty in changing systems.
@jessegrabowski@Dekermanjian -would you kindly recommend how I should go about creating my first PR or a beginner friendly issue I should focus on?
I can commit 12-17 hours/week
While I am enthusiastic about GSoC, I am equally interested in becoming a long-term contributor to the PyMC community beyond the program.
I apologize for the delay it’s been a rather busy two-weeks. The resources you’ve listed are excellent to learning Bayesian data analysis. I am going to defer to @jessegrabowski for details about GSoC, since it is my first time mentoring for this event.
When it comes to contributing to the PyMC ecosystem, my recommendation is to find issues that are labelled as beginner_friendlyand to read through the CONTRIBUTING.mdfile included in the repository.
Welcome to the community Hridayesh, looking forward to working with you!
Feel free to ping me about any questions you may have.
Thank you so much for responding @Dekermanjian. Nice to meet you too!
I have and will surely look at the “beginner_friendly“ issues and have gone through the “CONTRIBUTING.md“ file. I have also filed a documentation issue #8126 and tried to open a PR (first a darft and on recommendation of the maintainer and after a few changes another PR) #8127 and #8156.
Presently, I’m trying to learn more about the project’s topic and am trying to find exactly how it has been incorporated by PyMC and the community’s perspective on it.
I would really appreciate if you could point me to the correct files and documentation regarding this in the PyMC repo as I am having a slight problem locating them and would love to discuss the scope of the project with the mentors.
I have just submitted my Proposal for GSoC 2026 titled Scalable Sequential Inference for Bayesian State Space Models in PyMC.
I do realize that I should have done it earlier and have asked for feedback. I wanted to make sure that I have a jupyter notebook attached showing my familiarity with the pymc_experimental.statespace module but I faced countless errors, one after another, due to version mismatches and broken installation. After endlessly rebuiding my conda environment, I finally got all the packages to run properly and got to try out the module for myself. Notebook
This gave me no time to file a draft and ask for feedback. I deeply regret this and I hope it does not adversely affect my proposal. I have attached my Proposal here. @jessegrabowski@Dekermanjian I woukd be grateful if you could kindly go through it. Looking forward to any feedback ! GSoC.pdf (144.1 KB)
I’m sorry to hear you had issues with your environment. Please share any issues you run into on the pymc-extras GitHub repository. That way we can ensure that any version mismatches/broken installations don’t impact users.
Going through the videos and the example notebooks is an excellent way to familiarize yourself with the module. As you work your way through and build a custom SSM you’ll have a better grasp at the abstractions that the built-in models and structural module provide.
Great job putting together a proposal. I will review it soon. We hope you continue exploring and building with the PyMC ecosystem
Very happy to get your reply and insights. I will surely share the issues I faced in GitHub to ensure others do not face them.
Thank you for your reassurance. I am currently brainstorming as to what model I should build that would be best for me as a beginner and help me understand the library more deeply. Any tips and pointers would be greatly appreciated!
Finally, I hope you like my proposal (as silly and oddball it may be). Looking forward to continue working together irrespective of the results.
