I'm excited about the creative freedom and high impact of AI-driven software.

1.) In the past few months, I have architected and built a novel vision-based AI Agent and Framework using multimodal LLMs to autonomously interact (scroll/click) on websites and complete nontrivial tasks (a quick 1 min demo video is shown at the top of the page). [Read More]
The Framework portion consists of high-level operations (read, type, click_button, scroll, extract_info, etc.) that allows the AI to execute those actions and control the computer. The Framework uses Set-of-Marks Prompting, removing the need to use the mouse at all since all GUI interactions (including clicking buttons and scrolling) can be driven by the keyboard. The AI Agent portion automates dynamic tasks where the input can vary (unlike in traditional automation where actions are hard-coded). This capability is thanks to the vision-mechanism of the AI Agent and allows for "more human-like automation" since the Agent can see, interpret, and interact with website GUIs (just like humans). Beyond the proof-of-concept demo, there is a lot more engineering work (experiments, usecases, code development, etc.) and creative work (unqiue content creation & video demos, collaborations, community feedback, etc.) that can be done to take this project to the next level. Note, this project is very similar to Claude's Computer Use. However, some of the engineering design and implementation choices differ.
For more info, checkout the AI Agent Github Repo here.


2.) I created the first Open Source Youtube Comment & Associated Subreplies Scraper (No API key needed). [Read More]
AFAIK, no other open source Youtube Scraper scrapes and groups/labels Comments with their associated Subreplies for any Youtube video. The original inspiration for this project was to identify/report sophisticated yet subtle scam comment threads in Youtube videos. The future work consists of training or finetuning a ML model to identify those scam comment threads with high accuracy and then automate the reporting of those scams.


3.) Future(ish) projects in the works include Mad Money Summaries, [Read More]
SmileClips (create smile compilations from Youtube videos), Cracked Chess (chess but rules change dynamically, forcing players to make moves based on partial info), a better application of a well-known HCI device on laptops, and the Shortake App (a better alternative to the social media app BeReal).

I'm also interested in working with teams to build Physical AI. With a small team, I worked on a Wrist Wearable for Hand Gesture Recognition using real time Machine Learning. In a different embedded systems team project, we built an AI-powered Tic-Tac-Toe Robot that physically plays tic-tac-toe against a human at various difficulty modes using the Minimax algorithm.

In the past, I spent a summer as an Embedded Systems Intern at Intrepid Control Systems (a company making hardware & software for analyzing/testing onboard computer networks in vehicles). I supported firmware validation efforts by creating an ultrafast C++ GUI application for Ethernet Data Analysis and engineered 12 custom-made cables for internal testing purposes. During the school year, I co-taught one of Michigan's introductory engineering classes, E110, to 40 first-year students. For my final semester at Michigan, I was honored to be apart of TechLab at MCity, where I worked with a team to produce accurate 3D scenes from a set of 2D images using Gaussian Splatting in partnership with Capoom.

Outside of work, I try to get involved with my surrounding community. I have participated in multiple case competions and pitch events. My favorite was a Consulting Case Competition for P&G (Proctor and Gamble), where my teammate and I won 2nd place, providing actionable recommendations to increase P&G’s Hispanic shopper growth with a $12,000,000 budget. I was also the co-founder and Vice President of Michigan Forensics Speech and Debate Club, where I helped make the club website, recruit 150+ members, and hosted a grant-funded public speaking tournament.

Growing up, my two favorite sports were basketball and football. Looking forward, I hope to inspire and help younger generations experience the blessing of sports.