0.0 Start
I'm a second year at UC Irvine studying Data Science and Business Information Management. I like turning messy data into tools that people actually use, and I'm happiest when I'm building something, teaching something, or out on a long run.
Looking for Summer 2027 internships in AI, machine learning, and product management.
3.1 5K
I'd rather tell you exactly what my model can't do than oversell what it can.
My first real taste of machine learning came at a research camp in high school. That summer turned into a year of work and a published IEEE paper on network intrusion detection. I was 16. I still remember how good it felt to finally understand why a model was behaving the way it was.
Now I'm at UC Irvine, graduating in June 2028, studying data science on one side and business on the other. That mix is on purpose. I love the math, but what really gets me excited is the moment a model turns into something a real person can use. That's why I'm drawn to AI and machine learning work and to product management.
I grew up in the Bay Area, I go to school in Irvine, and after graduation I want to live in New York City. I'm always up for a good conversation, a hard problem, or a cup of coffee.
6.2 10K
Site Manager and Lead Instructor Intern
Machine Learning Intern
Research Assistant
13.1 Halfway
Now building · Solo
Predicts your marathon time and builds a training plan that adapts when life happens.
It learns from 2024 Boston Marathon results and has to beat Riegel's classic formula on runners it has never seen. It's still a work in progress, and I'm upfront that turning a training run into a race estimate is an assumption, not a proven result.
Python · scikit learn · pandas · Streamlit · pytest
Team project · Working demo
An AI weekly planner that turns Gmail and Google Calendar into one clear view of your week.
Every morning, Claude drafts three tone options for each email reply, so you just pick one and send. We built it as a team, and it's a working demo, not a product with users.
JavaScript · Anthropic API · Gemini API · Google OAuth · Netlify
Team of five · INF 43 Software Engineering
A sports matchmaking app for finding people to play pickup games with.
Our team of five built it for INF 43 with a React Native app and a Python REST API. I helped design the backend and its test suite, which reached 96% coverage with Jest and Supertest.
React Native · Expo · Python · FastAPI · Jest · Supertest
20.0 Mile 20
IEEE · 2024 Silicon Valley Cybersecurity Conference (SVCC) · Seoul, South Korea
I was 16 when this was published. It grew out of a high school research camp, and it was the first time I got to chase a question all the way to an answer.
Intrusion detection models often drown in features that are irrelevant or redundant. We asked a simple question: does the way you pick features actually matter? We tested four selection methods (entropy, mutual information, chi squared, and ANOVA) across two classic models and two neural networks on the NSL KDD dataset.
The takeaway: your choice of feature selection method matters a lot for classic models and much less for deep learning, which learns useful structure on its own. The work was supported by the National Science Foundation.
24.0 Mile 24
My goal is simple: run as many marathons as I can. Training is where I do my best thinking, and it's the reason Mile20 exists.
I'm a guard, and I was varsity captain at Saratoga High School. Captaining taught me more about leading people than any class has.
I've worked as a barista at BLK Dot Coffee. Ask me about espresso and I will talk for a while.
Ten years of trumpet. Through music events, I helped raise about $50,000 for Maui wildfire relief. These days you'll also find me at festivals and karaoke nights.
I'm in China every year and I snowboard Lake Tahoe whenever I can. Japan and Indonesia are next on the list.
26.2 Finish
I'd love to hear about internships, an idea you're working on, or your favorite race. Email is the fastest way to reach me.