Cyber threats are becoming increasingly common. Phishing websites, malicious links, fake messages, and social engineering attacks target millions of people every day.
Most security solutions rely heavily on cloud services: data is sent to remote servers, analyzed, and then a decision is returned to the user.
As someone who is passionate about both artificial intelligence and cybersecurity, I started asking myself a question:
Can we build a security assistant that uses AI directly on the user's smartphone while preserving privacy?
This question became the starting point of Aigris.
Who am I?
I'm a second-year Computer Engineering student with a strong interest in:
- Artificial Intelligence
- Cybersecurity
I enjoy building projects that allow me to combine theory with practical engineering challenges. Aigris is currently my biggest personal project and an opportunity to explore how AI and cybersecurity can work together in real-world applications.
What is Aigris?
Aigris is an experimental Android application that aims to become an AI-powered cybersecurity agent capable of helping users defend themselves against common digital threats.
The project focuses on a simple principle:
Security should be intelligent, proactive, and privacy-friendly.
The long-term vision is to create a mobile security assistant capable of:
- Detecting phishing attempts
- Analyzing suspicious URLs
- Identifying potentially dangerous content
- Warning users about digital threats in real time
- Performing as much analysis as possible directly on the device
Why On-Device AI?
Privacy is one of the main motivations behind this project.
Sending sensitive information to external servers is not always desirable. Modern smartphones are becoming increasingly powerful and can now execute lightweight machine learning models efficiently.
Running AI directly on the device offers several advantages:
- Better privacy
- Lower latency
- Reduced dependence on cloud services
- Some protection capabilities even while offline
- Greater user control over personal data
This privacy-first approach is one of the core ideas behind Aigris.
Technologies Behind the Project
Aigris combines multiple areas of computer science:
- Android Development
- Kotlin
- Machine Learning
- ONNX Runtime
- Cybersecurity
- Software Engineering
One of the most interesting aspects of this project is learning how to deploy AI models efficiently on mobile devices while respecting performance and battery constraints.
Current Goals
The project is still evolving, but the current areas of focus include:
- Phishing detection
- Suspicious URL analysis
- Real-time threat alerts
- Privacy-preserving AI inference
- Malware-related threat detection
Technical Challenges
Building an intelligent security application on a smartphone is not easy.
Some of the challenges I am currently exploring include:
- Optimizing AI models for mobile devices
- Reducing inference time
- Limiting memory consumption
- Balancing security and battery usage
- Designing a user experience that remains simple and non-intrusive
These challenges are exactly what make the project exciting from both an engineering and research perspective.
Future Ideas
Some ideas I would like to explore in future versions are:
- More advanced phishing classifiers
- Social engineering detection
- Better explainability of AI decisions
- Browser protection mechanisms
- Additional security modules
- Further optimization of on-device models
Why I'm Sharing This
One of the reasons I love open source is the opportunity to learn from other developers.
I'm still a student, and I know there is a lot I can improve. That's exactly why I'm sharing this project publicly.
I would genuinely appreciate feedback about:
- The architecture
- The idea itself
- The machine learning approach
- The Android implementation
- Possible security improvements
- Features that could make the project more useful
GitHub Repository
You can find the project here:
https://github.com/ullah-naqeeb/aigris.git
If you have suggestions, ideas, or constructive criticism, I would be extremely grateful to hear your thoughts.
Even a small piece of feedback can help improve the project and help me grow as a developer.
Thank you for reading!