An AI support center built for Ingress Academy students. Students can ask questions about courses, programs, and next steps in Azerbaijani, English, or Russian and receive answers based on the academy's own course materials.
Live and in use · AI assistant · Answers course-related questions · Helps students find the right course · Supports Azerbaijani, English, and Russian · Provides answers without waiting
The problem
Students ask the same questions about courses and what to do next, often in different languages. The challenge is not simply answering those questions — the answers need to come from the academy's own information. An assistant that relies on general AI knowledge can provide answers that are inaccurate or inconsistent with the academy's actual courses and approach.
What I built
An AI assistant powered by Ingress Academy's own course materials. Students can narrow their questions down to a specific area or course, ask questions in their preferred language, and return to previous conversations later.
Why it matters
Students can get answers about the academy's courses in their own language, instantly and without waiting for a staff member to become available.
My role
This is a real project I built for Ingress Academy. I developed the assistant end-to-end — from integrating the AI model and building the retrieval system that finds relevant information from the academy's course materials, to conversation persistence and the student-facing interface. The system is live and used by students every day. The student area is behind authentication, so there is no public demo link. The frontend visible to students is open source, while the backend and other system components are private.
Ingress AI · live
AI assistant
How it works
Student chat interface
Answers are generated from the academy's own course materials
AI model
Saved conversations
The AI assistant finds relevant information in the academy's course materials and uses it to form an answer, rather than relying solely on general knowledge. Conversations are stored, so a student can return to a previous discussion and continue later.
Built with: LLM API · Retrieval over course content · React · TypeScript · Vite · TanStack Query · REST APIs · Docker · nginx
