CyberDog | Devpost

Logo

Inspiration

With over 6 billion recorded cyber attacks in 2024, cybersecurity has become one of the most critical industries for keeping the internet safe. Yet despite its importance, the field faces an estimated 3.5 million unfilled jobs in 2025. One key barrier to entry is the need to pass information-dense certifications, which are often difficult to study for and retain. CyberDog was built to lower this barrier—helping aspiring professionals prepare for certifications faster, easier, and with deeper understanding.

What CyberDog does

CyberDog is a multi-agent learning app that transforms how people study for cybersecurity certifications. Based on the Feynman Technique—the idea that the best way to learn is by teaching—we let users play both student and teacher in an interactive dialogue with AI agents.

Currently, CyberDog supports CompTIA Security+, one of the most popular entry-level certifications. Our app centers around three key features:

Trainer (teacher agent)- This is an expert agent which a user can talk to, and can answer any questions they have about Security+.
Dog (student agent) - Acting as a naive student, this agent will ask the user random questions about Security+ material sourced from an AWS S3 bucket. The user's job is to recall the information they learned, and "teach" the student agent.
Student Assessment (Agent to Agent Communication): After the user teaches the Dog, the Dog consults the Trainer. The Trainer evaluates how well the Dog learned—essentially measuring how effectively the user taught, which directly reflects the user’s mastery of the subject.

What's next for CyberDog

Since we only cover one certification, what's next is expanding our coverage to more certifications to attract more users
We also plan to make our student and teacher framework more interactive by including reasoning for both student and teacher agents