Inspiration

Banks lose the most money not to outside hackers but to trusted people already inside: employees misusing privileged access, quiet data exfiltration, and slow manual investigations that find the damage too late. At the same time, quantum computing is putting today's encryption on a countdown. We wanted one platform that watches for insider threats the moment they start, responds on its own, and protects critical systems with cryptography built for the post-quantum era.

What it does

Nexus is an AI-powered cybersecurity platform built for the banking sector. It autonomously detects insider threats, monitors privileged access, runs behavioral analytics, and protects critical infrastructure using quantum-proof cryptography.

🤖 Guardian AI (NASE): the autonomous brain that analyzes activity and decides how to respond
🔐 Privileged Access Management: tracks and controls who can reach sensitive systems
🚨 Insider Threat Detection: spots abnormal behavior from trusted users in real time
📊 AI Risk Assessment: scores every user and action by risk
🛡️ Risk-Based Authentication: raises or lowers authentication requirements based on live risk
⚡ Autonomous Threat Response: contains attacks without waiting for a human
🔑 Quantum Security Vault: protects critical secrets with post-quantum cryptography
🎯 MITRE ATT&CK Mapping: links detected behavior to known attack techniques
🏦 Critical Banking Systems Monitoring: a live view of the systems that matter most
📄 Compliance & Audit Reports: generates investigation reports automatically

In the demo, Nexus detects an insider threat, calculates the risk, mitigates the attack, and generates an investigation report, all in real time.

How we built it

Frontend: React, TypeScript, and Vite for a fast, type-safe app
UI: Tailwind CSS and React Bits for a polished security-operations interface
Visualization: Three.js for the 3D threat view and Recharts for live risk analytics
AI: the Google Gemini API powers behavioral analysis, risk reasoning, and incident explanations
Reporting: jsPDF generates downloadable compliance and investigation reports

Challenges we ran into

Turning raw behavioral signals into risk scores that are both accurate and easy to explain
Making autonomous response feel trustworthy, with clear reasoning behind every action
Keeping real-time visuals smooth while running heavy 3D graphics and live charts
Building a complete, polished platform within the hackathon's time limit

Accomplishments that we're proud of

A full detect, assess, respond, and report loop working end to end
Mapping detected behavior to MITRE ATT&CK so analysts can act quickly
Bringing post-quantum security and AI-driven defense together in a single banking-focused platform

What we learned

We learned how much value there is in explainable AI for security: analysts trust a system far more when it shows why it acted. We also learned how to design for both speed and clarity under real-time conditions.