L.E.B.R.O.N. is a chatbot that uses a machine learning model to predict the outcome of NBA games and give the user advice on how to make favorable sports bets, narrated by none other than LeBron James himself.
demo.mp4
The chatbot front-end was built using the React javascript framework and the react-chatbot-kit package, connected to a LangChain conversational agent which called OpenAI's GPT-3 API. The LeBron voice narration was made using calls to the ElevenLabs text-to-voice API using a custom-made voice model. The chatbot runs locally.
The sports betting model was built and trained using TensorFlow, and we attempted to create an interface for it using FastAPI but ultimately didn't have enough time. Because the chatbot is powered by LangChain, it would be relatively straightforward to add the model as a tool for the conversational agent.
This project was completed in 24 hours for the HackAI 2023 hackathon with Suvel Muttreja and Wei-Chen Huang (Devpost link). I created the chatbot interface and the LangChain agent and integrated the voice narration, Suvel created the voice model and worked on the interface, and Wei-Chen trained and tested the betting model.
