Inspiration -

Cybersecurity threats are widespread and escalating. In 2024, over 422 million people in the U.S. were affected by cyber scams and attacks. We wanted to address this critical issue by building an app that can effectively detect and prevent these threats in real-time. We also have had some personal experiences in the past with this where we got scammed, and really wanted to make an impact and prevent other people from facing the same issues that we faced.

What it does -

SpamAssassin is a next-gen cybersecurity powerhouse. It scans emails for common threats like phishing, data breaches, and ransomware, offering users enhanced protection against online fraud with the use of Generative AI through Palantir AIP to maximize the accuracy.

How we built it -

  1. Building the backend
  2. Integrating it with the frontend
  3. Fine-tuning the AI model to do what was asked of it
  4. Bugfixes & Polishing it up

Challenges we ran into -

  • The AI helper often talks about irrelevant information, such as saying hello to the user, and wanting to be friends with the user. We were able to fix this by writing a more detailed prompt.
  • Instances where JavaScript unexpectedly converting variables to different types
  • Coordination - when the frontend encountered a setback and delayed, the backend had to follow along as well.

Accomplishments that we're proud of -

  • After countless challenges -- 4.5 hours of debugging, we successfully used good examples from our experiment to modify our prompt for the AI helper, teaching it how to properly select spam emails.
  • Using a new frontend framework that we were unfamiliar with
  • Coding and building in multiple different languages and platforms effectively and combing it together to maximize efficiency.
  • Being able to make use of Generative AI to effectively classify spam and fraud emails.

What we learned -

  • How to apply machine learning techniques to real-world cybersecurity problems.
  • The importance of user-friendly design, especially for critical apps like cybersecurity tools.
  • Effective teamwork and time management to meet tight deadlines and deliver results.
  • How to use the Palantir platform, so we did not have to rely on an internal backend.

What's next for Spam Assassin -

  • Cross-Platform Integration: Expand the app to mobile, browser, and desktop platforms for comprehensive device protection.
  • Social & Messaging Detection: Extend our app to detect scams across social media and messaging apps like Instagram, WhatsApp, and Discord.
  • Live Threat Alerts: Implement real-time notifications to alert users instantly about emerging scams and provide actionable security tips.

Built With

Share this project:

Updates