At Capsa AI, I assisted with improving Capsa's AI capabilities, with my main two projects being improving description accuracy on charts and figures, and improving row label matching for ingested data. I was able to gain experience working with both LLMs and traditional machine learning techniques, raising figure description accuracy to 90+% as well as cleaning table aggregation without noticeable time delay.

I assisted in MountainSeed’s R&D efforts into using LLMs to scan and answer questions based on appraisal reviews, testing different models and embeddings to build a prototype Retrieval Augmented Generation system and set the framework for future development. We were able to produce high-accuracy question answering models based on data points from our testing documents.

I worked for ten weeks with Total Systems, a company under Global Payments, as a summer intern for data analysis. Under the supervision of a Datalake manager, I learned to use Jupyterhub, Starburst, and dbt (Data Build Tool), allowing me to use SQL to quickly and efficiently transform data tables. At the end of my internship, I created a presentation on the last project I did where I used dbt to transform raw transaction tracking data into a more usable form.

The Georgia Governor's Office of Student Achievement runs a prestigious summer program for high school juniors and seniors called the Governor's Honors Program, or GHP, where students go through multiple rounds of nomination and application before being invited to a month-long summer program. I majored in math, taking the courses programming, problem solving, and combinatorial game theory. We also had the opportunity to make a final project related to our courses, and I, along with a partner, created a program using Processing that simulated the evolution of walking using genetic algorithms.

I led our school's VEX Robotics team, a high-school robotics competition where teams build robots that compete in a variety of tasks. I personally took roles in building, programming, and driving our robot. Our highest achievement was coming in second place at the state level.

Inspirit AI is a program taught by Stanford/MIT alumni and graduate students that introduces students to fundamental concepts of Artificial Intelligence. We learned how to use concepts such as multi-layer perceptrons and convolutional neural networks, culminating in a real-world project. My project was detecting emotions in faces.