Clive Dondi

🎯 Project Demo Hub

Welcome to my AI/ML project showcase. This site highlights my deployed machine learning demos, complete with working frontends, backends, and source code links.

---

📽️ 1. AIML - End-to-end Movie Recommendation System

A hybrid movie recommendation system built using content-based filtering and collaborative filtering, deployed as an interactive web application with Streamlit:

Key Features

• Customizable Hybrid Recommendations:
  • Adjust weights for content-based and collaborative filtering to get tailored results.
• Efficient Memory Usage:
  • Precomputed cosine similarity matrix stored in a FAISS vector store for content-based filtering.
  • Optimized SVD for collaborative filtering.
• Interactive Interface:
  • Built with Streamlit for a user-friendly experience.

🔗 Live demo: View App 💻 Source code: GitHub Repo

1. Enter a Movie Title:
   • Input a movie title (e.g., "Toy Story (1995)") to get recommendations.
2. Optionally Enter a User ID:
   • Provide a user ID to personalize recommendations using collaborative filtering.
3. Adjust Weights:
   • Use the sliders to control the balance between content-based and collaborative recommendations.
4. Set Number of Recommendations:
   • Choose how many recommendations to display.
5. View Results:
   • The app displays a list of recommended movies based on the inputs and settings.

---

📽️ 2. AIML - Telco Customer Churn Prediction System

This project is an end-to-end machine learning system to predict customer churn for a subscription-based business. The goal is to identify users who are likely to cancel their service soon so that retention teams can take proactive action.:

Key Features

• Tenure buckets and binary flags:
  • .
• One-hot encoding for categorical features:
  • .
• Feature interactions (tenure/charges):
  • .
• Resampling for class imbalance:
  • .
• Threshold tuning:

💻 Source code & README.md: GitHub Repo

---