DJ Bestie

Personalized AI DJ Page
Computer Vision Model in Action

Inspiration

We all crave a music buddy who not only shares our tastes but also vibes with our moods and conversations. Imagine having a DJ and a friend in one! DJ Bestie was inspired by the desire to bridge music discovery with real-time interaction, creating a personalized and fun experience for people who want to listen to tunes while feeling connected. With hands-free interactions and an animated character that moves and talks like a friend in the same room, DJ Bestie brings a uniquely personal touch to music discovery.

What it does

DJ Bestie is an AI-powered virtual DJ and companion. It chats with you, analyzes your facial expressions and body language, and curates playlists tailored to your mood and energy level. The hands-free design lets you talk to it naturally, just like you would with a close friend. DJ Bestie responds with real-time conversation, plays music, introduces similar songs and artists, and even reacts with a lively animated character that moves and talks with you. Whether you're chilling, working out, or partying, DJ Bestie is your go-to vibe curator.

How we built it

Front end: React Next.js Backend: Javascript, python, Typescripts, MongoDB, WebSockets

Challenges we ran into

Cross-Platform Integration
Recommendation System Optimization
Data Privacy Concerns
Real-Time Emotion Analysis
UI/UX Cohesion
Emotion Overlap Detection

Accomplishments that we're proud of

Innovative Multi-Modal Integration
AI-Driven Song Recommendation System
Seamless and responsive UI
Emotion-Sensitive Analysis
Mood-Based Playlists
Real-Time Performance and Efficiency

What we learned

Optimize Real-Time Processing
Overcoming AI model Limitations
Collaboration Under Pressure
User-Centric Design Principles
Training Multi-Model
Debugging Complex Systems
Edge Case Management

What's next for DJ Bestie

Integration with Popular Music Platforms
Improved Animations
Customizable Characters
Fitness Mode (for workouts, yoga, or dance sessions)
Offline Functionality
Social Sharing Features

Built With

deepface
keras
mongodb
next.js
openai
python
pytorch
react
spotify
tensorflow
typescript
unity

Submitted to

TAMUhack 2025
- Winner Third Overall Software Hack

Created by

Built the front end with React and Next.js. Had fun learning how to use Tailwind CSS.

John Mo
:D
I have contributed to the computer vision aspect, enabling real-time facial emotion recognition, body motion tracking, and user interaction analysis.

Ujjwal Gupta
Built our retrieval augmented generation (RAG) backend using Python, GPT, Spotify endpoints, and PyTorch. Designed our Unity WebGL interface with custom animations, lighting effects, and React integration.

Surya Jasper
I contributed to the backend development by building and managing the MongoDB database and integrating the Spotify API. My work focused on creating a seamless backend architecture to ensure personalized music recommendations and smooth interactions.

George Mathew