For my senior project I had the privilege of working with 4 good friends to create Dispatchr. Dispatchr aimed to provide “cloud based crowdsourced shopping”. The idea is that other students with cars can log onto Dispatchr when they are planning to go shopping, and pick up requests made by other users who live nearby and whose items can be found in the store that the student was planning on going to. Much like what we have now, UberEats, doordash, except community driven with rewards and compensation. Much of the frontend was create with React Native, allowing us to support both iOS and Android while iterating very quickly. The backend was a ruby on rails API server hosted on Heroku. We also used Dwolla to support ACH and credit card payments.
Inspired by Jian-Yang’s revolutionary Not Hotdog app, a group of four of us hopped on the food identification hype. The end result? An android app, aptly named SeeFood, capable of identifying any food you threw at it. The account system which allows users to access their valuable identified food on any device is powered by Firebase and the state of the art image recognition is a sophisticated proprietary algorithm. we actually used the clarifai api
During the summer of 2016, I had the amazing opportunity to intern at ViaSat Inc.. I collaborated with another UCSB student to design, develop and test software made to automate the company software release pipeline in a continuous integration/development environment. The original process would take approximately four hours to release a new software build. With the tools we built, that process was reduced to mere minutes. The engineer would run our command line tool or GUI tool, fill out a small number of fields, and the tool would take care of the rest. The command line tools were made so it can easily be integrated into continuous integration/development software (such as Jenkins, Travis CI, etc.), so releasing software would not require any action from the engineer. The GUI was built with JavaFX.
SpooderBot is an asynchronous, multipurpose chat bot for the VoIP service Discord. It’s built with Python 3.5 and implements several APIs to provide Discord members with music playing functionality (YouTube, Spotify) and image searches (GIPHY, Imgur). The bot also offers several administrative capabilities (point system, kicking/banning users, defining roles, etc) and debugging tools. Currently, SpooderBot is hosted on a inexpensive server running Ubuntu 16.04, used by two gaming communities with 50+ members, and its incredibly robust as it handles 20+ commands simultaneously with ease.
The following are some of the sites I’ve built over the years as an amateur web developer.