All projects

AI-Powered Customer Service & Appointment Booking System

Conversational booking system that pairs Twilio telephony with a fine-tuned LLaMA model for real-time, voice-driven appointment scheduling.

Python
Django
React
LLaMA
Twilio
MySQL

Overview

A customer-service system that answers calls, holds a natural conversation, and books appointments end-to-end without a human in the loop. Built around a fine-tuned LLaMA model wired into Twilio for telephony and a Django + React stack for the booking workflow.

What it does

  • Accepts inbound calls via Twilio and streams audio into the model layer
  • Generates dynamic, context-aware conversational responses using a fine-tuned LLaMA model
  • Books, reschedules, and cancels appointments against a MySQL-backed Django backend
  • Surfaces a React dashboard for staff to view, edit, and override bookings in real time

Stack

  • Voice & telephony — Twilio
  • Model — Fine-tuned LLaMA, served behind a Python service
  • Backend — Django + MySQL
  • Frontend — React

Why it mattered

Manual phone-based booking was the bottleneck. Replacing the first call leg with a conversational agent kept the customer experience warm while removing the back-and-forth that staff used to handle by hand.