Inspiration
Choosing what to watch can often be overwhelming due to the vast number of available movies. While many recommendation systems suggest similar titles, they rarely explain why those recommendations are relevant. We wanted to build a smarter movie discovery platform that combines the accuracy of Machine Learning with the conversational capabilities of Generative AI, making movie recommendations more interactive, personalized, and engaging.
What it does
MovieMind AI is an AI-powered movie recommendation platform that combines a content-based Machine Learning recommendation engine with Google Gemini AI. Users can search for movies, receive personalized recommendations, explore movies by genre, view movie posters fetched from TMDB, and get AI-generated explanations that make each recommendation easier to understand. The application provides a conversational experience instead of simply displaying a list of similar movies.
How we built it
We built MovieMind AI using Python and Streamlit for the application interface. The recommendation engine was developed using Scikit-learn, Pandas, and NumPy, implementing content-based filtering with cosine similarity to identify related movies. Google Gemini AI was integrated to generate conversational responses and recommendation explanations, while the TMDB API was used to fetch movie posters. The project was version-controlled with GitHub and deployed on Streamlit Cloud for easy public access.
Challenges we ran into
One of the biggest challenges was integrating an external AI service while ensuring the application remained reliable. Since AI-generated explanations depend on the Gemini API, we had to design the system so that the Machine Learning recommendation engine continues to function even if the AI service is temporarily unavailable or API limits are reached. We also worked on improving movie matching, handling conversational queries, refining genre recommendations, and creating a responsive, user-friendly interface.
Accomplishments that we're proud of
- Successfully combined Machine Learning and Generative AI into a single application.
- Built a conversational movie recommendation experience instead of a traditional recommendation list.
- Developed a reliable recommendation engine using content-based filtering.
- Integrated dynamic movie posters using the TMDB API.
- Designed the application to gracefully handle AI service unavailability while preserving its core functionality.
- Successfully deployed the project on Streamlit Cloud with a polished user interface.
What we learned
This project helped us better understand recommendation systems, cosine similarity, and content-based filtering. We gained practical experience integrating Large Language Models into a real-world application, managing external APIs, handling failures gracefully, deploying cloud applications, and designing AI-powered user experiences that remain reliable even when external services encounter limitations.
What's next for MovieMind AI
We plan to extend MovieMind AI by introducing hybrid recommendation algorithms that combine content-based and collaborative filtering, user authentication, personalized recommendation history, mood-based recommendations, watchlists, multilingual support, voice interaction, and user feedback mechanisms to continuously improve recommendation quality. Our long-term vision is to make MovieMind AI a comprehensive and intelligent personal movie companion.
Log in or sign up for Devpost to join the conversation.