Our drive and curiousity for machine learning, cyber security, and data science pushed us to make a complex antivirus with intruistic features beyond our years of experience by simply collaborating together and using the opportunity our parents gave us to make the best of ourselves.

Our antivirus was written mainly in python3 and is driven by supervised machine learning. We used Keras, and NumPy to supervise the AI. As an additional feature we used the classic signature method to help increase overall malware detection for our program. We used a dataset with PE(Portable Excutable Headers) to analyze/train our model to recognize malicious patterns. We also set up multiple monitoring systems for packet analysis, and RAM usage. We used data normalization to increase accuracy and learning speed for the AI.

One of the main challenges we ran into was formatting the .csv dataset the right way for the AI to start progressing properly. Our dataset was extremely big and wasn't formatted in the most efficient way that slowed our learning speed by hours.

- Having a full working AI model with 70% accuracy for malware detection
- Successfully learning how to use Keras, and Tensorflow library
- In-Depth profiency for detection of different malware features.
- Creating such a complex project in 24 hours
- Being completely sleep deprived for the whole competition 

- How to successfully implement machine learning libraries 
- Learned how to use supervised machine learning for malware detection
- Learned to use the Google Cloud Platform for AI development 

If things go right Knocking on some wood, we will continue to improve the accuracy, dataset, and marketing front for our program and become the next Elon Muskuteers. No pun intended.