Projects

If we assign happiness to certain rides and have information about how much time it takes to ride a string of rides in a certain order, how can we maximize happiness in one day at a theme park? This app (also embedded below) seeks to do exactly that. Here is the Github link. Demo Motivation There is a lot of information we can work with: Each ride has an associated happiness. Each ride has a wait time. Total time must not exceed the time the park is open. Walking between any two rides takes a certain amount of time. Throughout the day, wait times and walk times (crowd levels) change. My first thought was to use a maximum flow reduction with Ford-Fulkerson. However, there seemed to be too many factors which made Ford-Fulkerson seem infeasible. ...

This (Cornell GitHub) was our final project for CS 3110: Data Structures and Functional Programming. We implemented a simple transformer in OCaml and trained it on successful posts on SideChat, a social media app popular at Cornell and other schools. I worked with Haadi Khan, Domenic Fioravanti, and Will Bradley. Fig. 1. YouTube demo of our post generator made by Haadi. Probabilistic Model Before creating the transformer, we first created a probabilistic model. ...

This tool (GitHub) looks at de-vigorished sports betting odds for bets offered by Pinnacle, a sharp bookmaker, and uses those to find positive EV betting opportunities on FanDuel, a soft bookmaker. Tools like this are not unique, but they are offered as a service for high fees, so I decided to make a simple version myself. Rather than search for arbitrage opportunities, it looks for statistical arbitrage-esque market inefficiencies since sports betting “markets” are somewhat inefficient. ...