I study computer science and business at USC, and I spend most of my time at the seam between the two — building things, then figuring out how to get them in front of the people who need them.
Lately that's meant AI agent tooling, research systems, and product work. Before that, a nonprofit I started in high school.
My family moved a few miles down Rockville Pike and the school district around me changed completely — different resources, different support, different expectations for what students could do. YAFU started because none of that should turn on an address.
I built the platform from scratch: a web application where families could find resources and request help meeting basic needs, plus a second tool that streamlined volunteer recruitment for local shelters. Writing the code was never the hard part. Getting it in front of people was, which is how I ended up teaching myself SEO and social strategy.
From there it became tutoring — I interviewed tutors to run free virtual sessions — plus food drives, school pantries, and chapters run by students at nearby high schools.
200+ families
50+ tutor interviews
5 food drives
$2,000+ raised
YAFUA workshop on making masks out of recycled fabric, run during the pandemic.
Americans throw out around 200lbs of food a person each year, mostly on a guess. NoWaste lets you photograph something, describe what happened to it — "pizza left on a warm counter six hours, then refrigerated for eight" — and get a real read on whether it's still safe, from GPT-4 Vision and Llama 70B.
The decision I'm proudest of was refusing to ship only an app, because an app assumes a smartphone. So we built two things: a lightweight iOS app, and a Raspberry Pi unit with a camera and odor sensors in a housing we made on site, at roughly $100 to assemble, for communities where phones aren't a given.
On a team of four I owned product strategy, split the work by everyone's strengths, and pitched to the judges. It won Best Use of an API.
MobileThe iOS version, built in Swift.
RecognitionBest Use of an API — Postman, TreeHacks.
A backtesting engine for evaluating systematic strategies against market data — moving average, mean reversion, momentum — with the performance analytics you need to tell a real edge from an overfit one. More of what I'm working on lives on GitHub.
Experience
GoogleAI Software Engineer InternShipped a secure tool-use API widening a production browser agent's action space, and led the privacy redesign of its screenshot traces. New York.
Fox CorporationAI Product Manager InternBuilt an A/B testing framework across 1.2M+ monthly visitors that surfaced 46,000+ untapped trial starts.
AmazonQuality Assurance Engineer InternDesigned an internal AI tool that cut manual test planning time by 80% across the team.
AccentureAI/ML Technology Consultant InternLed five people building a summarization engine over 1,000+ news articles for client investment research.
USC Machine Learning CenterSoftware EngineerBuilt the lab's management system — the site researchers use day to day.
National Science FoundationResearch / Machine Learning InternRedesigned SPARQL workflows and Neo4j integrations, cutting data retrieval time by 75%.
National Institutes of HealthData Analytics InternAnalyzed 10+ billion genomic data points and co-authored two peer-reviewed publications.
KID MuseumEducatorSTEM workshops for elementary and middle schoolers — circuits, robots, code.
Off hours
Outside of work, I enjoy running, playing pickleball, trying new foods, and sitting in on an occasional poker game.
