AI Smart Assistant
Face Recognition | Voice Activation | GenAI-powered Responses
Developed by Jon Spahiu (an AI Passionate Student)
Scroll down to explore the project details.
Developed by Jon Spahiu (an AI Passionate Student)
Scroll down to explore the project details.
Welcome Hack Club visitors!
This AI Smart Assistant uses cutting-edge technologies including face recognition, speech-to-text, retrieval-augmented generation (RAG), and OpenAI’s GPT models to create a secure, intelligent assistant experience.
- Full Project Explanation
- Demonstration Video
- Presentation Slides
If you are part of the Hack Club, please follow and/or vote my project as I am trying to finish publishing it and hope for some shells to support this project!
Link to my project and the iframe is down below if you like to follow or/and vote my project!
Check out my github readme file below!
Link to my project and the iframe is down below if you like to follow or/and vote my project!
100% built in Python — the world's leading language for AI and machine learning projects.
| Step | Description |
|---|---|
| 1 | Detect known faces and activate the microphone securely |
| 2 | Convert user speech into text input |
| 3 | Query vector database (FAISS) using RAG for relevant information |
| 4 | Generate accurate responses using OpenAI's GPT models |
| 5 | Convert the response back into speech for the user |
| Component | Solution | Libraries/Tools |
|---|---|---|
| Face Recognition | Leveraged facial recognition libraries and integrated with a Pygame visualization window. | face_recognition, Pygame |
| Speech Recognition | Used Google's Speech Recognition API and pyttsx3 for text-to-speech output. | SpeechRecognition, Pyttsx3 |
| Vector Database | Used Embeddings to convert text to numbers which is sent to Vector Database to compare numbers for accurate results | LangChain, Vector, Embeddings |
| FAISS | Utilized FAISS for vector database queries to enhance response accuracy. | Meta FAISS, LangChain |
| Query and Response | Integrated OpenAI GPT models with RAG to produce context-aware answers. | OpenAI API, LangChain (LLM + RAG) |
| Multithreading | Implemented multithreading to ensure simultaneous process management. | Python Threading |
LLM (Large Language Model): AI models trained on massive datasets to generate human-like text (e.g., ChatGPT).
RAG (Retrieval-Augmented Generation): Combines LLMs with external databases to provide highly accurate, customized answers.
If you want to see the slides used in the Expo, you are in the right place!