hey, i'm rohan

I like making things I actually want to use.

I'm a computer science student at UVA, spending this summer in San Jose working on machine learning at Expedia. Most of my projects start with something I'm already into: running, movies, music, or a problem that keeps bugging me.

see what I've made say hey

right now

in: San Jose, CA
summer 2026: ML at Expedia Group
training for: my first Richmond Marathon

Rohan standing behind a waterfall in Iceland — somewhere along Iceland's Ring RoadMay 2026

a few favorites

Projects I kept coming back to.

see everything →

still working on it2026

Marathon Prep Bot

Solo builder and the runner using it every week

I wanted one place that could remember my training, put each workout in context, and help me tell whether I was actually improving. So I built it around my own workout history.

More than 1,300 workouts plus live iPhone strength logs now feed one coaching system, with encrypted private sync and a safe public dashboard.

typescriptdatasecurityfitness

take a look →

finished project2025

Shortlist

Built the retrieval model, Go serving system, and web app myself

Pick one to five movies you already love and Shortlist builds a fresh set of recommendations from the overlap in your taste. I trained that exact flow, built the Go API, and shipped the full product.

On 7,060 cold users held out by a global time cutoff and scored exactly once, the five-favorite flow reached 84.1% HitRate@10 and 0.331 Recall@100, against 73.8% and 0.228 for popularity on the identical users.

mlgopytorch

take a look →

still working on it2025

Today

Designed and built the native app, encrypted sync, and public/private data contract

Today is the iPhone app I use for quick morning weight, gym logging, and the day's coach plan. It stays useful offline, remembers my last sets, and feeds the right level of detail into the rest of my fitness system.

A signed iPhone app now links HealthKit, WorkoutKit, encrypted coach sync, and the live fitness dashboard without creating a second workout record.

swiftuihealthkitworkoutkit

take a look →

where I've worked

Some work I'm proud of.

I've gotten to work on machine learning, backend code, and tools used by real people. These are a few things I'm glad I got to build.

jun 2026 - present
Expedia Group
Machine Learning Engineer Intern
Built a FastAPI prototype for OneKey, a 100M+ member rewards program, with live profiles and safe fallbacks
may 2025 - jul 2025
Freddie Mac
Software Engineer Intern
Built Java Spring Boot microservices that process and validate 120k accounting events, deployed via Jenkins CI/CD
nov 2024 - may 2025
Biocomplexity Institute, UVA
AI/ML Research Assistant
Built scalable graph-analysis pipelines with cuGraph, SciPy, and NetworkX on 1.2M-edge ICIJ and crypto graphs

the full timeline →

the rest of me