Hi, I'm Jon! I'm an graduate researcher at Carnegie Mellon studying Automated Science. In my spare time, I enjoy pursuing computer science as a hobby. I have experience with Python, Go, MATLAB, and TensorFlow.

Here are some of the projects I am currently working on:

• D&DAV: An audio-visual dashboard for my D&D games, written in Python and using obsidian.md as a renderer.
• DiceCloud: An open-source, real-time character sheet for D&D 5e, written in Vue and JavaScript.
• BRCA-s: A capstone to my foray into Machine Learning, this uses a variety of ML methods (Such as PCA, K-means, Logistic Regression with L1/Lasso penalty, Random Forests) to find which differntially expressed genes are most significant when subtyping breast cancer. Employed cross-validation and hyperparameter optimization with GridSearch.
• MSAL: An Active Learning framework for training a NN surrogate model to MASSIF, a powerful but slow algorithm that finds the stress and strain fields of pourous materials. Uses Monte-Carlo dropout and COVDROP as measures of uncertainty, and a novel selection heuristic called USP that balances exploitation of model uncertainty with data structure.

• DegranQuant: A machine learning model to count activated granules in neutrophils, written in Python using TensorFlow and numpy.
• HYDRO: A closed-loop automated hydroponics system that balances ions in a nutrient maintank.
• Ten: A tetherless fluid handling robot for performing microbiology and molecular genetics experiments.

• LinkedIn: https://www.linkedin.com/in/jon-p-34ba2a129/

The best ways to reach me are via email (jonathanpotter18 at gmail dot com) or Discord (@Jonpot).

I am currently a full-time student committed to introducing Active Learning loops into Automated Biology Labs in Pittsburgh.