Loan Defenders Documentation

Welcome to the Loan Defenders documentation! This project showcases Microsoft Agent Framework and Azure AI Foundry through an intelligent multi-agent loan processing system that delivers decisions in under 2 minutes.

🔒 Private Repository with Public Documentation

Last Updated: October 2025

Documentation: Always publicly accessible at this site

Source Code: Private repository - request access to explore implementation

Requesting Access:

Send a LinkedIn message to Nikhil Sachdeva with:
- Your GitHub username
- Reason for interest (learning, research, evaluation, etc.)
- Brief background on your multi-agent AI interests
Access typically granted within 2-3 business days
Access provided to researchers, students, and developers interested in multi-agent architectures

Public Resources (no access needed):

✅ Complete documentation and guides (this site)
✅ Architecture Decision Records (ADRs)
✅ System architecture diagrams
✅ Development philosophy and patterns
✅ Deployment guides and best practices

🦸 What is Loan Defenders?

Loan Defenders is a demonstration of multi-agent AI orchestration where specialized AI agents work together to process loan applications. Think of it as your AI Dream Team for financial services:

🦸‍♂️ Cap-ital America - Loan Orchestrator coordinating the entire workflow
🦸‍♀️ Scarlet Witch-Credit - Credit Specialist analyzing creditworthiness
🦸 Hawk-Income - Income Specialist verifying employment and income
🦹‍♂️ Doctor Strange-Risk - Risk Advisor synthesizing all assessments

✨ Key Features

🤖 Microsoft Agent Framework

5 specialized agents with distinct personalities and expertise
Autonomous decision-making powered by Azure AI
Real-time coordination through agent threads
Tool integration via MCP (Model Context Protocol) servers

⚡ Lightning Fast Processing

< 2 minute end-to-end processing time
No forms to fill - conversational interface
Real-time updates as agents process your application
Complete transparency - see every agent's assessment

🎨 Modern Tech Stack

Frontend: React 19 + TypeScript + Vite
Backend: Python 3.12 + FastAPI + Microsoft Agent Framework
Infrastructure: Azure Container Apps + Azure AI Foundry
Developer Tools: UV package manager + MkDocs Material

📢 Get Started with the Code

Request Repository Access

The source code repository is private. Documentation remains publicly accessible, but repository access must be requested.

How to Get Access:

Send a LinkedIn message to Nikhil Sachdeva with:
- Your GitHub username
- Reason for interest (learning, research, evaluation)
- Brief background on your multi-agent AI interests
Wait for approval (typically 2-3 business days)
Clone the repository once access is granted

Once You Have Access:

# Clone the repository
git clone https://github.com/niksacdev/loan-defenders.git
cd loan-defenders

# Create a feature branch
git checkout -b feature/your-feature-name

# Make your changes and submit a PR

Benefits of Repository Access:

✅ Full source code access for learning
✅ Submit pull requests for improvements
✅ Open issues for bugs or questions
✅ Join discussions about architecture and design

See our Contributing Guide for detailed instructions.

🚀 Quick Start

New to Loan Defenders? Start here:

Getting Started

Learn about the architecture and installation process

Quick Start Guide
Request Access

Request repository access and start building with the code

Installation Guide
Product Guide

Learn about the revolutionary loan experience and business strategy

Product Documentation
Architecture

Understand the multi-agent system design and technical decisions

System Architecture
Deployment

Deploy to Azure with Container Apps and CI/CD pipelines

Deployment Guide
AI Development

Learn about the AI-augmented development philosophy and workflow

Development Philosophy

📸 See It In Action

Landing Page - Meet Your AI Dream Team

The home page introduces the AI Dream Team of specialized loan processing agents

Live Agent Workflow - Watch the Magic Happen

Real-time agent coordination as each specialist processes your application

Instant Results - Loan Approved! 🎉

Complete loan decision with full transparency and agent recommendations

🏗️ Architecture Highlights

Multi-Agent Workflow

graph LR
    User[👤 User] --> Coordinator[🦸‍♂️ Coordinator]
    Coordinator --> Intake[📋 Intake Agent]
    Intake --> Credit[💳 Credit Agent]
    Credit --> Income[💰 Income Agent]
    Income --> Risk[⚖️ Risk Agent]
    Risk --> Decision[✅ Decision]
    Decision --> User

MCP Tool Servers

Application Verification (Port 8010) - Validates loan application data
Document Processing (Port 8011) - Processes uploaded documents
Financial Calculations (Port 8012) - Performs complex financial calculations

Azure Deployment

Azure Container Apps for API and UI
Azure AI Foundry for agent orchestration
Azure Application Gateway for load balancing
Azure Application Insights for observability

🎯 Why This Project?

Loan Defenders demonstrates several cutting-edge concepts:

Multi-Agent Orchestration - Complex workflows coordinated by autonomous agents
Personality-Driven AI - Each agent has unique personality and expertise
Tool Integration - MCP servers provide agents with specialized capabilities
Production-Ready - Full Azure deployment with observability and monitoring
AI-Augmented Development - Built by 1 developer + AI agents (productivity of 8-12 person team)

📚 Documentation Sections

Explore the complete documentation:

Getting Started - Installation, setup, and first loan application
Product Guide - Product vision, strategy, and user experience design
Architecture - System design, agents, and technical decisions
Deployment - Azure deployment, CI/CD, and infrastructure
AI Development - AI-augmented development workflow and philosophy

🤝 Contributing

This private repository welcomes contributions from authorized collaborators. The project showcases AI-augmented development and multi-agent architecture patterns.

How to Contribute:

Request repository access via LinkedIn message to Nikhil Sachdeva
Clone and explore the codebase once access is granted
Submit pull requests for improvements
Open issues for bugs or feature requests
Join discussions about architecture and design

See our Contributing Guide for detailed instructions.

📞 Support

Support Policy - Private Repository

Status: Private repository with active maintenance

Support Channels:

Documentation: Always publicly accessible (this site)
Issues: Available to collaborators with repository access
Pull Requests: Welcome from authorized collaborators
Discussions: Available for architecture questions (requires access)
Security Issues: Report via GitHub Security Advisories or LinkedIn message

Response Times:

Access requests: 2-3 business days
Issues/PRs: Best effort, typically within 1 week
Security issues: Prioritized, within 48 hours

Resources:

Documentation: You're reading it! 📖
Discussions: Ask architecture questions (requires access)
Using This Repo: Guide for users (requires access)

🎓 Learning Resources

Microsoft Technologies

Microsoft Agent Framework GitHub - Multi-agent orchestration framework
Microsoft Agent Framework Documentation - Official framework documentation
Azure AI Foundry - Enterprise AI platform and services
Model Context Protocol - Tool integration specification

Project-Specific

Development Philosophy - AI-augmented development approach
AI Workflow - How this project was built
Design Principles - Ethical AI and development principles

Ready to Get Started?

Install the project and run your first loan application in 10 minutes!

🚀 Quick Start Guide