Personalized Content Generation for Bank of Baroda Hackathon 2024

Team: Coach_Sahab

"Driving the technology through the leaps of Himalayas"

Project Overview

This project aims to solve the problem of personalized content generation for the banking sector, enhancing customer satisfaction and engagement through tailored content. Our solution leverages advanced AI and machine learning models to deliver personalized marketing materials, financial reports, and educational content.

App Download

Link: Click Here

Demo Video

Link: Click Here

Screenshots

AI Financial Advisor	Dashboard Home Screen	AI Generated Financial Report
Personalized AI Bots	Personalized AI Bots	Bot Detector
AI Based Smart Notification	Profile Screen	Graph and Goals
Login Screen	Splash Screen	AI Policy Recommender
AI Loan Recommender	Transaction	Payment Screen

Repository

• Android Code Branch Link: SmartBaroda
• GenAI Backend Branch : GenAI Services
• Core Backend Branch : Core Server

Architecture

Overall Architecture

1. Data Ingestion Layer

   • Components: Azure Data Factory, Azure Data Lake Storage
   • Function: Collects and processes customer data from multiple sources, ensuring data quality and consistency.

2. AI Model Layer

   • Components: Azure Machine Learning, Azure Kubernetes Service (AKS), Azure Batch
   • Function: Utilizes AI models to generate personalized content based on customer data analysis.

3. Distribution Layer

   • Components: Azure API Management, Azure Functions, Azure Logic Apps
   • Function: Distributes personalized content through appropriate channels (email, SMS, mobile app).

4. Feedback Loop

   • Components: Azure Application Insights, Power BI
   • Function: Gathers feedback on content effectiveness to refine and improve personalization algorithms.

Overall Architecture (For the final production ready application)

Backend Architecture

1. Source Code Management

   • Tools: GitHub, Jenkins/GitHub Actions
   • Function: Version control, continuous integration, and deployment.

2. Deployment

   • Components: Cloud-based virtual machines, Docker
   • Function: Hosting the application and running it in containers.

3. AI System

   • Components: AI models for processing complex queries
   • Function: Provide advanced functionalities.

4. API

   • Function: Interfaces for communication between software components, handling client requests.

5. Database Management

   • Components: MongoDB
   • Function: Storing and retrieving application data efficiently.

Backend Architecture Diagram

Generative AI Architecture

1. Langchain Agent

   • Function: Uses a language model to choose a sequence of actions to take.

2. Langchain Tools

   • Function: Interfaces that the agent can use to interact with the world.

3. OpenAI LLM (Azure)

   • Components: OpenAI ChatGPT 3.5/4.0, Langchain Framework
   • Function: Generative AI for creating personalized content.

Generative AI Architecture Diagram

Frontend Architecture

1. Retrofit Service

   • Function: Simplifies network operations, manages API calls, and converts responses into data models.

2. DI Class for Dagger

   • Function: Manages dependencies, ensuring clean and reusable components.

3. ViewModel

   • Function: Manages UI data and business logic.

4. Repository

   • Function: Provides a single source of truth for data, handles caching and synchronization.

5. Activity/Fragment

   • Function: Displays data, handles user interactions, and delegates logic to ViewModel.

Android Architecture Diagram

Azure Resources Required

• Azure AI Studio & Azure OpenAI: For advanced language models.
• **Azure App Services **: For Deployement of the Backend for core and generative AI services

Methodology

1. Pilot Testing
2. Feedback and Refinement
3. Gradual Rollout
4. Marketing and Awareness Campaigns
5. Customer Support and Training
6. Continuous Improvement

Key Differentiators

• Highly Personalized Content
• Real-Time Content Generation
• Comprehensive Integration
• Data-Driven Insights
• Continuous Improvement

Adoption Plan

1. Pilot Testing
2. Feedback and Refinement
3. Gradual Rollout
4. Marketing and Awareness Campaigns
5. Customer Support and Training
6. Continuous Improvement

Scalability

• Cloud Infrastructure
• Microservices Architecture
• AI Model Deployment
• Data Processing Efficiency
• Distribution Channels
• Continuous Monitoring and Optimization

Security Considerations

• Data Encryption
• Access Control
• Compliance Certifications
• Network Security
• Threat Detection
• Data Residency
• Audits and Assessments
• Disaster Recovery

Android App Releases

Version History

v1.0.0 - July 31, 2024

• Initial release with core features
• Includes functionalities like user registration, login, GenAI Features etc.

Contributors

• Yash Kamal Saxena (Generative AI Infra Developer)
• Tushar Garg (Backend Core Infra Developer)
• Uphar Gaur (Android Infra Developer)

Thank you for considering our project!