Designed and implemented an AI-powered review reply automation system to help businesses manage and personalize responses to Google reviews at scale.

• Integrated with Google Business Profile API to fetch real-time customer reviews across multiple locations and businesses.
• Engineered a customized reply generation pipeline using LLMs (Llama via Ollama + LangChain) with RAG (Retrieval Augmented Generation) to provide highly personalized, warm, and context-aware responses.
• Built logic for language detection and multilingual support, automatically generating replies in the same language as the original review (ex: English, French).
• Developed customizable business profiles and templates, allowing replies to adapt to each business’s tone, style, and industry nuances.
• Implemented advanced logic for review sentiment detection:
• Positive reviews → auto-posted with friendly replies
• Neutral/negative reviews → sent to approval dashboard for manual verification before posting
• Built web-based management interface (Firebase + FastAPI + Firestore) to:
• Onboard new businesses
• Manage review replies and approvals
• Generate monthly PDF reports using ReportLab, summarizing review trends, sentiment analysis, and AI replies
• Supported multi-location businesses with multi-profile management and per-customer customization.