Multi-Input College Chatbot Development

I am building an AI-powered assistant for our campus that can understand students and staff no matter how they reach out—typed messages, spoken questions, or even photos of notice boards and forms. The core job is to create the full conversational system, wire it to our existing data sources, and make sure it remembers context across an entire back-and-forth so users never feel they are starting over. What the bot must handle • Course information, timetables, and key dates • Event details and real-time updates • A searchable staff directory with contact methods • Personal student info that sits behind login (e.g., enrollment status, fees) Essential capabilities • Accept text, convert voice to text (and optionally reply with TTS), and extract text from images via OCR so a photo of a timetable or poster is enough to trigger the right answer. • Natural-language understanding fine-tuned for academic terminology. • A memory layer that lets the conversation stay coherent over multiple interactions, whether a student returns an hour or a week later. • Clear fallback or escalation when the bot is unsure. Deliverables 1. Working chatbot hosted on our preferred cloud platform (AWS or GCP is fine). 2. Source code with clean documentation and environment setup instructions. 3. Admin dashboard for updating FAQs, course data, and events without redeploying. 4. Short video demo that shows text, voice, and image queries in action. 5. Handover session so my IT team understands maintenance and future tuning. Acceptance will be based on the bot correctly answering a test suite of mixed-format queries, maintaining context over at least three consecutive turns, and deploying without errors in our staging environment. If you have experience combining NLP, speech-to-text, and OCR—and enjoy crafting smooth conversational flows—let’s get this rolling.

Project ID: 40374539