Research
AI Market Entry Strategy
We apply reinforcement learning (RL) to market entry and exit strategies, demonstrating that RL agents outperform rule-based approaches in maximizing capital. Our findings highlight RL's potential to uncover novel insights in microeconomic strategy.
AI Codenames
We use AI bots, vector embeddings, and cosine distance to demonstrate AI's effectiveness in the board game Codenames. A Spymaster leverages learned embeddings to generate clues, guiding both human and bot players in identifying their team's spies.
Unsupervised Polish Handwriting Recognition
We leverage an HWR model as a generator in a GAN network to predict Polish text from segmented image snippets. With no labels for these image snippets, we hope to show that the generator learns to model the distribution of real Polish text downloaded from the internet, enabling robust handwriting recognition without labeled training data.
Personal Projects
Gallery Friends
A simple web app I made back in the summer of 2023 that hosts user's profiles and pictures in a simplified social media format. I wrote the both the frontend and backend in NextJS, the site is hosted on Vercel, and the database is MongoDB.
RaicesUSA Website
A company website I created for my wife's Ecuadorian folklore dance group Nuestras Raices. I recently put this together and enjoyed making the site look nice and professional. It helped teach me a lot more that I didn't previously know about UX design and CSS.
AI SQL Project
Although this was part of some coursework, I played around more than I needed to with the idea of taking natural language questions (NLQs), passing them through ChatGPT with openai tokens, having the llm output an SQL command I could then run a query with, and using the result of a query as part of a prompt for the llm to respond to in a user-friendly natural language way.
Accomplishments
ICPC 2024 Hexadeciblue Team
In 2024, I competed in the International Collegiate Programming Contest (ICPC) for the Rocky Mountain Region. Our team placed 8th in the region and 2nd of all BYU teams who attended. This was an amazing accomplishment and it really taught me how to work with a team to problem solve. This is where my love for problem-solving and competitive programming really took off even more.
BYU Dean's List
In my last semester as an undergraduate student at Brigham Young University (BYU), I got notified that I was added to my college's Dean's list. This means that I ranked within the top 5% of my college for my last semester.
BYU CPC Advisor
Since Fall of 2024, I have served as one of the BYU Competitive Programming Club's (CPC) advisors. I have had the opportunity to serve other students and help them learn algorithms and have led the beginnings of some meetings. I have also been engaged in service opportunities like volunteering for career fairs to earn money for the club to fund its activities.
Other Projects (Work Related, Etc)
Relative Finder
As an undergraduate researcher, I had the privilege of working on a rewrite for an old project called Relative Finder. This project syncs with data from FamilySearch and allows you to see how you are related to deceased famous people, other people that you are interested in seeing your relationship to, and other live people if you invite them to use the app. This application boasts the usage of a little over 1,100,000 users and has on average 3,000 active users per week. I mainly worked on connecting the frontend to the backend, fixing backend bugs, and helping with the app's deployment through Terraform and GitHub Actions.
Reverse Indexing
Reverse Indexing is an application in which users are given a batch of up to 12 images at a time accompanied with a transcription that was predicted for the particular batch of images by a handwriting recognition model. Users simply mark images that appear to be incorrectly transcribed as incorrect, and then move on to the next batch. This was used in our handwriting recognition pipeline by allowing human-in-the-loop validation to help us improve our handwriting recognition model. I mainly worked on the frontend and did a couple tweaks to the backend.
Other BYU Family History Technology Lab Apps
I worked a little bit here and there on every other deployed app on the BYU Family History Technology Lab's website. My contributions to the rest of these applications are not as significant as the aforementioned projects; however, I gained more experience in just about everything in the fullstack developer experience including writing Terraform code and using GitHub Actions to deploy and manage every single application seen on our lab's website.