LeratoTheko/RSL-CRM-Chatbot

Development of Generative AI for Customer Relationship Management system for Revenue Services Lesotho # Generative AI Customer Relationship Management (CRM) Chatbot **Overview** Welcome to my Generative AI CRM chatbot project, designed as part of my data science portfolio. This chatbot is built to assist Revenue Services Lesotho (RSL) by answering frequently asked questions, enhancing customer interaction through AI-driven natural language processing. The chatbot uses GPT-2, fine-tuned with RSL's FAQ data, to generate responses. The project not only demonstrates my ability to work with advanced machine learning models but also highlights my skills in full-stack development, as I have implemented the user interface using Django. **Project Goals** Create a Generative AI CRM system for RSL that can handle user inquiries efficiently. Build an intuitive Django-based interface where users can interact with the chatbot. Showcase my data science and machine learning expertise by fine-tuning GPT-2 to generate responses from the FAQ data. Store, process, and evaluate large-scale language models to simulate a conversational agent for customer relationship management. **Key Features** GPT-2 Fine-Tuning: The GPT-2 model has been fine-tuned on RSL's frequently asked questions to provide contextually relevant responses. Django Frontend: A clean and intuitive interface built with Django, where users can interact with the chatbot. Error Handling: Input validation and basic error handling to ensure the chatbot remains functional, even with misspelled or incomplete user inputs. Generative AI Expertise: This project demonstrates my understanding of natural language processing and generative models, which are crucial skills in the field of data science. **Tech Stack** Machine Learning Model: GPT-2 (fine-tuned) Backend Framework: Django (Python) Frontend: HTML, CSS, JavaScript (for chatbot interface) Deployment: (Planned for deployment on Render/Heroku) Model Training: TensorFlow and Hugging Face transformers library **Installation** **1. Clone Repository** git clone gith …