Posts by Abhipsha Das
Career and life goals, basically
- 31 August 2021
Thinking about my future career is as daunting to me as any person in their 20s - laden with anxiety, uncertainty and still with a lot of optimism. Most of the times the anxiety part of it wins though and I am left with a lot of hours of overthinking. And the fact that applications and contacting people are such stressful processes hardly helps the matter.
I had always wanted to have a career in academia and pursue research, and have from the beginning of my undergrad been interested in machine learning. It is a vast field in itself and has so many cross-discipline applications that it sometimes leaves me confused since I have no idea what I would like the most to study. This made me wonder about taking up jobs or internships to rake up some experience and gain more of a nuanced understanding of the industry, but while they have been invaluable experiences, I am nowhere near sorted out in my head.
Modifying My Expectations
- 12 August 2021
It is perhaps one of my personality traits to not be able to follow any timeline that I have set up for myself, and this time is no different. I could easily become a raging cynic here but that would be too painful and boring to read, so I will refrain from that since I do not want to scare you away!
My original timeline for the Outreachy internship with PyMC3 looked like this:
Commits, Commits and More Commits
- 30 July 2021
It’s been a few months since I began contributing to pymc-examples and the experience so far has been enlightening. For someone who previously could only covetously look upon my fellow classmates as they contributed away to open source projects while I was still struggling with git and the jargon used in a project, I feel like I have come a relatively long way, thanks in no small part to the PyMC community members. I have gained some understanding of Bayesian inference and statistical concepts while updating the example notebooks, which are a set of jupyter notebooks hosted for the reference of newcomers trying to do probabilistic programming with PyMC3.
For some context, Bayesian inference means drawing statistical inferences using Bayes’ theorem to update the probability of a premise/hypothesis as and when we obtain more evidence for our defined premise. We define a statistical model for our data based on which a likelihood function and prior probability is derived, and based on these two, a posterior probability is obtained. It is a common and convenient way to update our hypothesis inside a statistical framework. So the steps of Bayesian inference can be summarized as: get observed data, build a probabilistic model to represent the data, specify prior distributions and apply Bayes’ theorem for inference and updating our hypothesis.
Everyone Struggles, But Everyone Believes They Are The Only One Struggling
- 18 July 2021
While I have had a very curious interest in FOSS since almost the beginning of my time at university, and tried to follow it up by looking at projects that intrigued me, I was never able to contribute to anything largely due to my inability to ask (read: fear of asking) questions like what a beginner like me could do. I often did not know how to structure my thoughts and questions in a way that the maintainers understood - beyond the initial “hey! I am a beginner in xyz and would love to work on this project” - because what I probably wished for is someone to assign me a task and list out all the prerequisites I’d need to fulfill it.
Since anyone reading this will know that that is not how open source exactly works - not that maintainers don’t help you out, quite the opposite, but delegating unresolved issues to newcomers like a teacher would do to their students is not how it goes. Any contributor needs to be more proactive in claiming issues, recognizing bugs, and approaching the project as a whole. Now what should people who really do want to be proactive in such a manner but actually feel lost and clueless do? I was one of them, and I can say that I continue to be like that. When I began my journey with Outreachy through the contribution period, after making sure that I had some interest in the PyMC project, I claimed issues without thinking much at all. I was encouraged by my mentor to scan through the many many open issues and chose the ones that seemed the simplest, and haven’t stopped learning and working since.
Keeping It Simple
- 10 June 2021
It should come as no surprise to anyone that writing about oneself is a conspicuously difficult thing to do no matter how facile it may seem on surface. Yet by a stroke of (incredibly good) luck here I am, doing a self-introduction online despite my unreasonable aversion to being perceived, as ludicrous as that may sound.
There are many things, however, that I could say about myself. I am easily distracted, largely indecisive and interested in pursuing too many varied things. I have a propensity to vacillate between having an outburst of inititative and doing nothing at all. I prefer cold weather to hot and I really hope to be able to play Paganini’s Caprices on the violin one day. I do not understand a thing about web development and am endlessly fascinated (but woefully underinformed) by artificial intelligence and its applications. I love learning languages and learning about languages, and I am perpetually trying to understand my relative position and privilege in the world. I would like to read more history, politics and visit every corner of the world, and feel happy to just be - without having to attach the worth of my life to what I accomplish or don’t accomplish.