I've consolidated some of my python projects over the years. Some projects are much older than others, so my coding style (and quality) has changed significantly throughout.
Here's a short explanation of each directory, organized into my own ad-hoc groupings
This is a python project to ingest abstracts from pubmed, and then fine-tune/test LLMs to try to detect databases as apart of the text.
This is full pipeline project where I collected and cleaned data, then fed it through a bunch of machine learning models to predict the outcome of basketball games. Was super proud of the resulting dataset I produced, and the final analysis.
This is another pipeline ingests multiple novels, analyzes a given theme, and produces a comparative book report with citations.
This was an interesting project I recently came across and wanted to share as well. In the course, we played around with different LLM methodologies in the domain of Law. Some interesting aspects of SMT and SAT solvers, including training llms to produce these texts. One of my first exposures to Boolean satisfiability problem, so it was very interesting. All of the code is in ipynb files.
Tracer was more of a learning project for playing around and debugging ai agents. It helped me learn alot more about how agents work and are orchestrated together. Essentially, it lets developers record and debug agent interactions with external services.
A fastAPI backend for generating websites using LLMs.