I built OmniBot AI, a full-stack dashboard designed to simulate and test automated WhatsApp Business workflows powered by AI. It bridges the gap between raw data storage and real-time LLM grounding.

Key implementation details: • Knowledge Base Engine: Implemented a system context pipeline that allows users to write dynamic business rules/FAQs and sync them directly into vector context for the bot. • Bilingual AI Support: Integrated Gemini AI with explicit prompt routing to automatically detect, process, and reply in both English and Arabic. • Live MongoDB History Stream: Built a real-time JSON log viewer on the frontend that streams database updates, tracking session language, token usage, and cache success states. • Infrastructure Analytics: Developed a monitoring component to track total synchronization logs and calculate average API response latency (optimized down to 1.2s).

Tech Stack: MongoDB, Express.js, React, Node.js, and Google Gemini API.