01

Project Overview

Revolutionizing blockchain security with AI-powered forensics

🎯

Scope

AltFlex is an integrated AI and digital forensics framework designed to proactively detect and analyze security exploits in cross-chain bridges and DeFi protocols. The system combines machine learning anomaly detection with blockchain forensic analysis to provide comprehensive security solutions for the Web3 ecosystem.

⚡

Complexity Assessment

  • Technical Complexity: High (AI/ML, Blockchain, Full-stack)
  • Architecture: Microservices (FastAPI + Next.js)
  • Data: Real-world exploit datasets, smart contracts
  • Integrations: Etherscan API, XGBoost ML

Expected Deliverables

Phase 1
✅ Completed

Flash Loan Detection

ML anomaly detection system

Nov 17 - Dec 31, 2025
Phase 2
✅ Completed

Address Verification

5-layer security pipeline

Jan 1 - Jan 31, 2026
Phase 3
✅ Completed

Premium Frontend

Enterprise-grade Next.js UI

Feb 1 - Feb 19, 2026
Final
📦 2027

Production Release

Complete security framework

Next Year
02

Agile Scrum Methodology

Why iterative development is essential for this project

01

Iterative Development Required

  • AI/ML models need continuous refinement based on test results
  • Frontend UI requires iterative design improvements
  • Security features demand rapid response to vulnerabilities
02

Complex, Evolving Requirements

  • Blockchain security landscape changes rapidly
  • New exploit patterns emerge requiring quick adaptation
  • External API integrations may require adjustments
03

Cross-Functional Skillsets

  • Backend development (Python, FastAPI)
  • Frontend development (TypeScript, Next.js)
  • Machine Learning (XGBoost, Feature Engineering)
  • Security Engineering (Rate limiting, Auth)
04

Demonstrable Progress

  • Working increments every 2 weeks for stakeholders
  • Sprint reviews provide accountability
  • GitHub discussions enable transparent collaboration

Why Not Waterfall?

❌ ML model accuracy unknown upfront
❌ Cannot predict all security requirements at start
✅ Agile allows pivoting when exploit patterns change
✅ Continuous testing catches security flaws early
03

Project Timeline & Sprints

10 sprints delivering 182 story points

Phase 1: Foundation & Core Detection

November 17 - December 31, 2025

Sprint 1 Week 1-2

Data Collection & Preparation

  • Etherscan API integration
  • Dataset creation (5 real exploits, 50 transactions)
📁 data/ directory with validated exploit samples
Sprint 2 Week 2-3

ML Model Development

  • XGBoost classifier training
  • Feature engineering (22 features)
🤖 AnomalyDetector with 100% validation accuracy
Sprint 3 Week 3-4

Rule Engine & API

  • 6 flash loan detection rules
  • FastAPI endpoint implementation
🌐 12 REST API endpoints with Swagger docs

Phase 2: Security Enhancement

January 1 - January 31, 2026

Sprint 4 Week 5-6

Address Validation

  • 5-layer verification pipeline
  • Format, checksum, ICAP, on-chain, behavioral
🔐 AddressVerifier module with external APIs
Sprint 5 Week 6-7

Behavioral Analysis

  • Velocity scoring & funding pattern analysis
  • Blacklist integration
📊 BehavioralAnalyzer with OFAC/mixer detection
Sprint 6 Week 7-8

API Hardening

  • Rate limiting & API key auth
  • Request validation & audit logging
🛡️ SecurityMiddleware with 100 req/min limit

Phase 3: Premium Frontend

February 1 - February 19, 2026

Sprint 7 Days 1-2

Landing & Dashboard

  • Next.js 14 setup, Tailwind CSS theme
  • Landing page with animations
🎨 Responsive landing + dashboard layout
Sprint 8 Days 3-4

Transaction Analysis UI

  • Transaction input form
  • Risk score visualization
📈 /dashboard/analyze with ML integration
Sprint 9 Days 5-6

Verification & Exploit UI

  • Address verification interface
  • Exploit database browser
🔍 /verify and /exploits pages
Sprint 10 Days 7-8

UI Polish & Deployment

  • 1900px max-width optimization
  • Documentation update
🚀 Production-ready codebase on GitHub

Sprint Velocity & Performance

182 Story Points Delivered
100% Completion Rate
18-22 Points per Sprint
10 Total Sprints
20
18
22
20
19
21
15
16
17
14
S1S2S3S4S5 S6S7S8S9S10
04

Team & Roles

A cross-functional team driving innovation

👨‍💻

Jay Arre P. Talosig

Technical Lead & Backend Developer
  • Overall architecture design & decision-making
  • Backend API development (FastAPI)
  • ML model training & feature engineering
  • Code reviews & GitHub management
  • Sprint planning & backlog prioritization
🎨

Alexander Castilo, Mark Jhoshua Taberna

COM231 - Frontend Developer
  • Next.js frontend implementation
  • UI/UX design & responsiveness testing
  • Component library development
  • API client integration
  • Browser compatibility & QA
🧪

Nicko Nehcterg Dalida, Rinoah Venedict Dela Rama

COM232 - QA Engineer & Data Analyst
  • Test suite development (pytest)
  • Exploit dataset curation & validation
  • Performance testing & security audits
  • Documentation writing
  • Sprint retrospectives

Scrum Ceremonies

📅

Daily Standups

Async via GitHub Discussions

15 min posts
📋

Sprint Planning

Start of each 2-week sprint

1-2 hours
🎬

Sprint Review

Demo working features

30-45 min
🔄

Sprint Retrospective

Continuous improvement

30 min
05

Technical Architecture

Production-ready microservices design

flowchart LR subgraph Data["📊 Data Layer"] A[Etherscan API] --> B[EtherscanCollector] C[Sample Data] --> D[data/] end subgraph Processing["⚙️ Processing"] B --> E[FeatureEngineer] D --> E E --> F[AnomalyDetector] E --> G[ExploitDetector] E --> H[BehavioralAnalyzer] end subgraph API["🌐 API Layer"] F --> I[FastAPI :8000] G --> I H --> I end subgraph Frontend["🖥️ Frontend"] I --> J[Next.js :3000] J --> K[Landing Page] J --> L[Transaction Analysis] J --> M[Address Verification] J --> N[Exploit Database] end style I fill:#6366F1,stroke:#4F46E5,color:#fff style J fill:#10B981,stroke:#059669,color:#fff

Project Flow Diagram

graph TB subgraph P1["Phase 1: Foundation"] A[Sprint 1: Data Collection] --> B[Sprint 2: ML Training] B --> C[Sprint 3: API Development] end subgraph P2["Phase 2: Security"] C --> D[Sprint 4: Address Validation] D --> E[Sprint 5: Behavioral Analysis] E --> F[Sprint 6: API Hardening] end subgraph P3["Phase 3: Frontend"] F --> G[Sprint 7: Landing & Dashboard] G --> H[Sprint 8: Analysis UI] H --> I[Sprint 9: Verification & Exploit UI] I --> J[Sprint 10: Polish & Deploy] end J --> K[Final Submission] style A fill:#6366F1,stroke:#4F46E5,color:#fff style B fill:#6366F1,stroke:#4F46E5,color:#fff style C fill:#6366F1,stroke:#4F46E5,color:#fff style D fill:#10B981,stroke:#059669,color:#fff style E fill:#10B981,stroke:#059669,color:#fff style F fill:#10B981,stroke:#059669,color:#fff style G fill:#8B5CF6,stroke:#6D28D9,color:#fff style H fill:#8B5CF6,stroke:#6D28D9,color:#fff style I fill:#8B5CF6,stroke:#6D28D9,color:#fff style J fill:#8B5CF6,stroke:#6D28D9,color:#fff style K fill:#F59E0B,stroke:#D97706,color:#fff

Development Tools

Backend Stack

Python 3.10+ FastAPI Uvicorn XGBoost scikit-learn NumPy Pandas pytest

Frontend Stack

Next.js 14 TypeScript Tailwind CSS 4 Framer Motion Lucide React

DevOps

Git GitHub Virtual Env npm
06

Tools & Artifacts

Project management and quality assurance

GitHub Repository

Version control, code hosting, issue tracking

View Repository →
💬

GitHub Discussions

Async standups, sprint planning notes

View Discussions →
📋

GitHub Projects

Kanban board for sprint backlog

View Board →
📝

Documentation

README, CODE_REVIEW, specifications

View Docs →

Definition of Done (DoD)

✅ Code peer-reviewed and merged to main
✅ Unit tests written and passing (pytest)
✅ Documentation updated (README, CODE_REVIEW)
✅ Feature manually tested in browser/API
✅ No critical security vulnerabilities
07

Risk Management

Identified risks and mitigation strategies

Risk Probability Impact Mitigation Strategy
🌐 External API downtime (Etherscan) Medium High Implement caching, fallback to sample data, retry logic
🤖 ML model low accuracy Medium High Iterative feature engineering, cross-validation, ensemble methods
📱 UI responsiveness issues Low Medium Test on multiple screen sizes, use Tailwind responsive utilities
📈 Scope creep High Medium Strict sprint goals, "nice-to-have" vs "must-have" prioritization
👥 Team member unavailability Medium Medium Documentation-first culture, cross-training on stack components

Quality Assurance Strategy

159 Unit Tests (pytest)
API Integration Tests
UI Manual Testing
🔒 Security Audits
08

Success Criteria

Measurable outcomes across all phases

Phase 1 Metrics

✅ Achieved
  • ✅ 6 flash loan detection rules implemented
  • ✅ ML classifier with 100% validation accuracy
  • ✅ 12 REST API endpoints functional
  • ✅ 23 unit tests passing

Phase 2 Metrics

✅ Achieved
  • ✅ 5-layer address verification operational
  • ✅ Behavioral analysis (velocity, funding)
  • ✅ API security: rate limiting, auth, validation
  • ✅ 123 additional tests (146 total)

Phase 3 Metrics

✅ Achieved
  • ✅ Premium Next.js UI with 5 pages
  • ✅ Responsive design (1080px-1920px)
  • ✅ Real-time health monitoring
  • ✅ 1900px max-width optimization
09

Reference Links

Project resources and documentation

AltFlex GitHub Repository

Main codebase hosting version control, issue tracking, and project documentation

→
💬

Agile Scrum Sprint Discussions

Async standup updates, sprint planning notes, and team collaboration

→
🔍

Phase 2 Development Code Review

Detailed technical review of address verification and security enhancements

→
📋

Capstone Methodology Design

Methodology design document for the AltFlex capstone project

→
📊

AltFlex Sprint Planning & Task Board

Interactive sprint backlog and task tracking board for Agile Scrum workflow management

→

Conclusion

The Agile Scrum methodology has proven highly effective for the AltFlex capstone project. The iterative approach allowed for rapid adaptation to emerging blockchain security requirements, continuous stakeholder feedback integration, and demonstrable progress through working increments every 2 weeks.

Key Success Factors

🔄
Flexibility

Pivoted from Streamlit to Next.js based on design requirements

👁️
Transparency

GitHub-based workflow provides full visibility to instructor

✅
Quality

100% test pass rate maintained across all sprints

🤝
Collaboration

Async communication via Discussions enables flexible teamwork

The project successfully delivered a production-ready AI-powered forensic framework through disciplined sprint execution, comprehensive testing, and continuous integration practices.