AlekoGPT is a Retrieval-Augmented Generation (RAG) chatbot built using Hugging Face Transformers and Dash. It allows users to ask questions about Alexandros Frangiadoulis' experience, skills, and background.
The chatbot loads personal knowledge from a structured markdown file, retrieves relevant context using semantic similarity (FAISS + Sentence Transformers), and generates contextual answers using the flan-t5-base model from Hugging Face.
Note: The model used is an open, free, and lightweight language model provided by Hugging Face. While it performs well for general understanding, it may have limitations in accuracy or consistency, especially when handling ambiguous queries or complex language structures.
The app runs locally using Dash, with a clean Bootstrap interface and is ideal as a personal AI assistant portfolio piece.
Visit the GitHub repository to view the source code of AlekoGPT.
Explore the model on Hugging Face used to power the chatbot.