Hi, I'm Kaushik Moudgalya 👋

I am a Data Scientist. I like code, coding, and NLP related to code.

I learn and play with transformers, Deep Learning math, LLMs, and Python. I write about these things as well, head on to my Website or Medium if you'd like to read what I've written. I also do traditional Data Science during my 9-5. I'm trying to learn more MLOPs nowadays but my one true love will always be code based models, whether they be LLMs or not.

Check out my recent work, TASTY: A Transformer based Approach to Space and Time Complexity.

I am working on the next version of the paper above! Exciting things await!

kaushik = {
  interests             : ["Code based LLMs", "Arithmetic Reasoning with LLMs", "Gamifying and accelerating Learning"]
  code                  : ["Python", "MATLAB", "VBA", "CSS?", "Java?", "Javascript?"] 
  skills                : ["Deep Learning", "NLP", "Shell Scripting", "Latex", "Machine Learning"]
  more_skills           : ["SQL", "HTML", "Git", "AWS", "GCP", "Robotic Process Automation", "Docker", "CSS?"]
  ask_me_about          : ["Transformers", "Machine Learning", "Manga", "Anime"]
  currently_learning    : ["Docker", "Vim", "K8s", "Vue"]
  fun_fact              : ["I can speak 7 languages"]
}
# ? marks denote acquaintance and non-expertise

Etrama's Projects

airflow-docker

My copy of the source code below, incorporating local changes needed:Source code of the Apache Airflow Tutorial for Beginners on YouTube Channel Coder2j (https://www.youtube.com/c/coder2j)

anime-ml-project-old

We'll do some ML in this project related to Anime. Starting with Anime / Non-Anime classification, moving onto Anime Recognition, then to Character Recognition.

cs230-proj

Deep learning for options pricing.

data_structures_practice

Practicing some Data Structures and Algorithms.

dl-paper-notes

My notes on a few DL Papers.

econometrics-with-python

Tutorials of econometrics featuring Python programming. This is a crash course for reviewing the most important concepts and techniques of basic econometrics, the theories are presented lightly without hustles of derivation and Python codes are straightforward.