Back to projects
Sohbatgah AI
Full Stack Developer & AI Engineer
Company
Sohbatgah Project
Period
08.2024 — 01.2025
Overview
An advanced conversational AI platform with multimodal capabilities, featuring secure authentication and intelligent context retrieval using RAG and vector search.
Challenge
Creating a conversational platform that could understand context, maintain conversation history, and provide accurate responses while ensuring data security.
Solution
Built a Spring Boot backend with Gemini AI integration, implemented RAG with PGVector for semantic search, and secured authentication with Spring Security. Created an intuitive frontend that handles multimodal interactions.
Key Highlights
▪Implemented RAG pipeline for context-aware responses
▪Built multimodal conversation flows with text and speech
▪Designed secure user authentication and data encryption
▪Optimized vector search for low-latency responses
▪Created admin dashboard for monitoring and analytics
▪Deployed to production with monitoring and alerting
Technologies
JavaSpring BootSpring SecurityGemini AIPostgreSQLPGVectorRAGVector SearchReactTypeScript