Skip to content

Projects

Things I built from problems I noticed.

Most of these started the same way — something bugged me enough to build the fix. Filter by what kind of work it was, or by the domain it lives in.

Type
Domain

10 projects

PersonalComing soon

BoxIt

As an out-of-state student, I saw college storage was expensive, inconvenient, and low-trust — and that students would rather earn money hosting it.

Systems/BackendFull-StackInfra/Cloud
Personal

Liberation Day and After

I'm fascinated by how huge macro movements ripple through sectors and reshape market structure.

QuantDataApplied-Research
Personal

EquiTable

I noticed Atlanta's food-insecurity problem was driven by disconnected food banks and a lack of understanding — I wanted to bridge that gap.

ML/AIFull-StackInfra/Cloud
Hackathon

FirstWave

Emergency demand is predictable — Friday nights in the Bronx, summer weekends in Brooklyn — but ambulances still sit and wait for the call instead of moving ahead of it.

Full-StackFrontendDatadraft
Personal

Benchwarmer

I've always been into the NBA and into predictions, and I wanted to bring real quantitative reasoning to it — not gut calls, but a backtested engine meeting live markets.

QuantML/AIData
Research

Steering Truth in LLMs

The standard ways to find a 'truth direction' in a model — linear probes, contrastive mean differences — leave signal on the table, so I tried to feature-engineer better directions and actually steer on them.

ML/AIApplied-Researchdraft
Personal

jobmaxxing

Internship postings are scattered across a dozen boards and go stale fast — I was tired of refreshing all of them, so I built one feed that watches them for me.

Systems/BackendDataInfra/Clouddraft
Personal

IMC Prosperity 4

I wanted to see whether my read on a market would actually hold up against a live order book — not just a clean backtest.

QuantDatadraft
Hackathon

clearRx

Whether two prescriptions interact shouldn't come down to a provider's memory — but a tool that answers it is only useful if you can trust the answer.

ML/AIFull-Stackdraft
Research

Aquatic Sustainability

Clean water fails quietly in the places with the least data — I wanted a way to see watershed risk and water-access gaps on a map before they turn into emergencies.

ML/AIDataApplied-Researchdraft