The first year of my PhD studies had, in all honesty, some rather unexpected challenges. Given the ongoing pandemic, my move to a new city for my PhD studies was put on hold. Instead, I decided to spend more quality time with one of my good friends (he’s a great cook!) and my parents. As spring was looming in, I quickly ran out of motivation to work on assignments and write my end-of-term report, but, with the help of my support network and all things considered, my first year ended well!
With summer right around the corner, I was eager for some good news and an opportunity to look forward to. On Monday, May 17 right before 2pm, I received the following email and nearly floored with happiness:
and I immediately shared this amazing news with my parents and my friends. I am extremely grateful for the Google Summer of Code (GSoC) opportunity for the many reasons. Firstly, I will learn more about open source development and probabilistic programming from well-established mentors. Secondly, this practical opportunity would complement well my summer studying for my PhD comprehensive exams. Last but not least, I’m being paid!
Under the mentorship of Christopher Fonnesbeck and Austin Rochford, my GSoC project centers around extending the PyMC3 package with a Dirichlet process submodule. PyMC3 is a probabilistic framework in Python that allows users to fit Bayesian models via MCMC sampling. There is a growing interest in Bayesian nonparametric methods for fitting statistical models without specifying its parametric form. Dirichlet processes can be used as priors for probability distributions themselves. How exactly? I shall find out in the next three weeks of community bonding!
In the coming months, I look forward to getting to know the PyMC3 community, reading more about Bayesian nonparametric statistics and, above all, learning from my mentors. Stay tuned for more!