AI-Powered Visual Search & Inspiration Matching Tool

This is NOT a chatbot project. We are specifically looking for an AI engineer with experience in computer vision, semantic image search, and multimodal AI. We are an established home décor export company in India with approximately 50,000 product images (15 GB). The images are already organized into folders by product category and named using SKU numbers. We need an internal desktop/web application. Core Feature 1 Index our existing product image folders. Allow natural language search. Examples: Black aluminium tray Organic bowl Modern candle holder Japandi decor Coastal accessories Core Feature 2 (Most Important) Upload an inspiration PDF (buyer catalog, trend book, mood board). The AI should analyze the PDF and retrieve the most visually similar products from our own catalog. Requirements: Keep our existing folder structure ( open to changes only after being communicated about them ) No ERP required. No inventory management. No website. Fast local search. Clicking a result should open the original product image. Preferred Technologies: Python Computer Vision CLIP / SigLIP or equivalent image embeddings Vector Database (Qdrant or ChromaDB) OpenAI or Gemini APIs where appropriate FastAPI (or similar) When applying Please answer: Have you built semantic image search before? Have you built inspiration/image similarity search? What technical approach would you recommend? Can you show a demo?