🌍

My Projects

my.uscis.gov

This is the website I work on in my current position, contact me offline if you would like to hear details!

Link

CRISPR Genomic Edit Detection

Building off research by M. Heath Farris et. al, I designed and developed an ensemble model involving an LSTM Autoencoder and statistical test to detect the presence of CRISPR edits.

I used python, (tensorflow, keras, numpy, pandas, seaborn), bash scripts, MITRE's High Performance Computing Cloud system, and ggplot2 in R for this research.

The abstract for the poster can be found on page 1489 in the following link:

Link

Liars Dice Agents and Backrooms

One of my favorite games to play with friends is Liar's Dice (a.k.a Perrudo, Dudo). Liar's Dice is a fascinating game involving luck and deception, and I've been obsessed with systems underlying the game. I've done some statistical analysis of the game using R and have coded the following agents in Python to simulate the game: Random Agent (Always chooses randomly), Probabilistic Agent (Always chooses action that maximizes probability) and a Neural Network Agent that acts based on estimating reward for actions using a nerual network.


I have also created a Liar's Dice Backrooms webapp(Inspired by Infinite Backrooms) where Large Language Models play against each other with a variety of prompted strategies and personalities.

I am currently working on a website and paper with my analysis - I'll link them here when I have them!

For now, play Liar's Dice online with your friends on my friend's website liars-dice.net that I am working to implement my AI's with!

Link

Statistical Analysis of the Card Game Rugby

I did an analysis of the card game rugby for my Computational Statistics class in Grad School. Rugby is a card game where cards are arranged face down in a diamond and a player goes row by row selecting one card from the row trying to get to the end. If the player flips a face card, they start over.

The project consists of analysis of Monte Carlo simulations of the game to determine if there is an optimal strategy. The answer is: there is! (sometimes)

The following link has a powerpoint with the findings:

Link