Scam Defender
A browser extension that detects if the user is visiting a phishing website using machine learning and displays a warning if they are. This browser extension is hooked up with an API hosted on Google Cloud that uses a Keras model trained on the UNIMAS dataset to determine if the site is a scam.
Inspiration
Every day, thousands of people are taken advantage of by internet scammers. Our team's goal was to help prevent this.
How we built it
We started off by creating a design doc with all teammate agreement. We then created a trello board to keep track of tasks. We split up the duties mostly between front end, data science, and DevOps relatively evenly.
What we learned
DevOps is hard. Machine learning is hard. Never use Tensorflow.js under any circumstances.
Challenges we faced
We had a lot of trouble getting the model good enough to actually be meaningful because a significant amount of our data was not in English and we had to filter this out. We also had a lot of problems hosting our API. For whatever reason, post requests stopped working for us at around 5am so we had to remake the API to use get requests.
Log in or sign up for Devpost to join the conversation.