Menufesto: Agentic RAG Dish Translator

Overview

Developed a sophisticated translation application that uses Large Language Models and Retrieval Augmented Generation (RAG) to provide culturally accurate descriptions for Chinese dishes.

Key highlights

• Designed an agentic RAG pipeline with query rewriting and a reflection module to evaluate context quality.
• Implemented a Gemini Search fallback mechanism for high-precision factual grounding.
• Architected the system using Docker-based microservices to support modular scalability and reproducible development.