🍿 Movies MH24

A modern movie exploration app focused on smart search, semantic recommendations, and efficient large-scale data handling

22,000+ movies | MongoDB Atlas Search | Vector embeddings | No paid embedding APIs

🔗 movies.mahfuzul.vercel.app

✨ Key Features

Extensive Movie Database: Contains 22,000+ movies with detailed metadata for browsing and filtering.
Advanced Filtering: Users can filter movies by genre, year, and type with multi-level filtering options.
Intelligent Search: Leveraging MongoDB Atlas Search, the platform supports typo-tolerant and relevance-ranked searches.
Semantic Recommendations: Using vector embeddings generated with Nomic EmbedText v1, the platform provides users with movie suggestions based on similarity in plot and content.
Cost-Efficient Embedding Pipeline: Embeddings for all 22,000 movies were generated using a self-hosted model on Google Colab to avoid third-party API costs.
Dynamic Recommendation Engine: When visiting a movie page, users see related movies computed through vector similarity, enhancing discovery beyond simple genre matching.
Responsive & Interactive UI: Built with Next.js, TailwindCSS, Radix UI, and Framer Motion for smooth, modern user experience.

🛠️ Technical Highlights

Frontend: Next.js with TypeScript, responsive design, server-side rendering.
Backend: MongoDB with Mongoose, advanced search indexes, and API routes for filtering and recommendations.
Machine Learning: Vector embeddings for semantic similarity using Nomic EmbedText v1.
Data Handling: Efficiently managing and embedding a large dataset of 22,000 movies while enabling fast, real-time recommendations.

📈 What I Learned / Demonstrated

Hands-on experience with large-scale data handling and database optimization.
Practical understanding of search indexing and relevance scoring with MongoDB Atlas Search.
Applied knowledge of vector embeddings and semantic similarity for recommendation systems.
Gained experience in integrating modern frontend frameworks with backend APIs for interactive web applications.

🔗 Links

✨ Key Features

Extensive Movie Database: Contains 22,000+ movies with detailed metadata for browsing and filtering.

Advanced Filtering: Users can filter movies by genre, year, and type with multi-level filtering options.

Intelligent Search: Leveraging MongoDB Atlas Search, the platform supports typo-tolerant and relevance-ranked searches.

Semantic Recommendations: Using vector embeddings generated with Nomic EmbedText v1, the platform provides users with movie suggestions based on similarity in plot and content.

Cost-Efficient Embedding Pipeline: Embeddings for all 22,000 movies were generated using a self-hosted model on Google Colab to avoid third-party API costs.

Dynamic Recommendation Engine: When visiting a movie page, users see related movies computed through vector similarity, enhancing discovery beyond simple genre matching.

Responsive & Interactive UI: Built with Next.js, TailwindCSS, Radix UI, and Framer Motion for smooth, modern user experience.

🛠️ Technical Highlights

Frontend: Next.js with TypeScript, responsive design, server-side rendering.

Backend: MongoDB with Mongoose, advanced search indexes, and API routes for filtering and recommendations.

Machine Learning: Vector embeddings for semantic similarity using Nomic EmbedText v1.

Data Handling: Efficiently managing and embedding a large dataset of 22,000 movies while enabling fast, real-time recommendations.

📈 What I Learned / Demonstrated

Hands-on experience with large-scale data handling and database optimization.

Practical understanding of search indexing and relevance scoring with MongoDB Atlas Search.

Applied knowledge of vector embeddings and semantic similarity for recommendation systems.

Gained experience in integrating modern frontend frameworks with backend APIs for interactive web applications.