I built an ML-powered FAQ Chatbot that understands user intent using NLP. 🔧 Tech Stack: • Python 3.14 + Gradio 5.x for interactive UI • Sentence-Transformers (all-MiniLM-L6-v2) for semantic embeddings • PyTorch for model inference • Cosine Similarity for intent matching 💡 Key Learning: Instead of keyword matching, the bot converts questions to 384-dimensional vectors. This helps it understand rephrased queries and improves accuracy by 70%. 🎯 Challenges Solved: 1. Handled Gradio 5.x breaking changes with `type="messages"` format 2. Implemented threshold-based fallback for out-of-scope questions 3. Deployed with Git version control on GitHub The complete code is open-sourced here: 👉 https://github.com/Sachin-SDE12/Chatbot-Codealpha-Task-2-