multi-tenant AI customer support chatbot

Built a multi-tenant AI customer support chatbot using Next.js, Express, PostgreSQL, and Prisma,

handling tenant isolation and full-stack architecture.

• Designed and implemented a RAG pipeline with document chunking, embeddings (Nomic), and similarity

search via pgvector, using Redis + BullMQ for asynchronous processing.

• Integrated LLM inference (Ollama, Gemma4) and real-time chat, enabling context-aware responses from

user-provided knowledge bases.