Fweb Smart Platform

Project Summary

Full-stack intelligent web platform built on CodeIgniter 3
Integrated Flask-based AI recommendation system and Role-based dashboards for tailored user experiences
Payment integration for seamless transactions
Geolocation-based personalization to enhance user relevance
Push notifications for real-time user engagement
Scalable and dynamic architecture for performance and growth

Technologies

Frontend:

HTML, CSS, Bootstrap, JavaScript

Backend Framework:

CodeIgniter 3 (Core Application)

Flask (AI Microservice for Recommendations)

Database:

MySQL

AI & Data Science:

Neural Collaborative Filtering (NCF)

Hybrid Filtering (Content + Click-based)

Python (NumPy, Pandas, TensorFlow/Keras)

Challenges 😨

Integrating AI recommendations built in Flask with the CodeIgniter 3 backend.
Ensuring real-time FCM push notifications work seamlessly with geolocation-based targeting.
Managing multiple payment gateways while maintaining a secure and smooth user experience.
Scaling Neural Collaborative Filtering for large datasets with high accuracy.

Solution ✅

Developed a REST API for communication between Flask AI service and CodeIgniter 3.
Implemented JWT-based authentication for secure API calls.
Integrated geolocation services to refine AI recommendations and target notifications.
Optimized NCF model training using efficient batch processing and caching results.

Research Plan

The development roadmap for Fweb includes several strategic enhancements to elevate its intelligence and user experience. It begins with AI enhancement, aiming to improve recommendation accuracy using transformer-based models and context-aware personalization. Next, real-time capabilities will be introduced through WebSocket-based updates for notifications, product changes, and order tracking. The platform will then be expanded into a Progressive Web App (PWA), offering offline support and enhanced push notification functionality. Finally, a focus on data-driven optimization will involve collecting usage analytics to fine-tune AI models and improve geolocation-based targeting for more personalized user engagement.