We are building a curated modest fashion marketplace called Mizyana, and we are looking for an experienced AI developer to help us build an AI-powered product catalog standardization agent. Our sellers provide product data in many different formats such as Excel files, Google Sheets, PDFs, or directly from their websites. We need an AI system that can ingest these sources, extract the product data, normalize it into our structured template, and rewrite product descriptions to match our marketplace format and tone of voice. The goal is to automate a large portion of the repetitive work involved in onboarding sellers and preparing product listings for the marketplace. This is not a simple chatbot project. We are building a structured data pipeline that combines extraction, AI transformation, validation, and export. Project Scope The system should: 1. Accept seller product data from multiple sources: • Excel or CSV files • Google Sheets exports • PDF product catalogs • Seller websites (crawl product pages) 2. Extract product information such as: • Product title • Description • Sizes • Colors • Materials • Variants • Images • Price 3. Map extracted fields into our standard marketplace schema. 4. Normalize product attributes such as: • color naming • size formatting • category classification • material naming 5. Use AI to: • rewrite product titles • rewrite descriptions • standardize product copy • apply our marketplace tone of voice 6. Run validation checks such as: • missing required fields • conflicting attributes • possible duplicate products • low confidence extraction 7. Send low-confidence rows to a human review queue. 8. Export final data into: • Excel template • CSV • JSON / API payload ⸻ Example Workflow Seller uploads Excel file or submits website URL. System flow: Seller Input → Data Extraction → Field Mapping → Data Normalization → AI Content Rewrite → QA Validation → Human Review → Export to Marketplace Catalog ⸻ Preferred Technology Stack We are open to suggestions, but the current preferred direction is: Frontend React or Next.js Backend Python (FastAPI preferred) AI Layer OpenAI API Extraction Firecrawl for websites Python tools for spreadsheets OCR or document parsing for PDFs Database PostgreSQL Workflow orchestration Optional: n8n, Temporal, or queue-based processing ⸻ Deliverables 1. Functional ingestion pipeline 2. AI mapping and rewriting logic 3. QA validation layer 4. Human review interface 5. Export system 6. Documentation ⸻ Nice to Have Experience with: • ecommerce product catalogs • marketplace platforms • PIM systems • LLM structured outputs • web scraping • data pipelines • AI agents ⸻ Project Type Initial build with possibility of ongoing work. We are building a long-term marketplace platform and expect additional AI tools after this phase. ⸻ To Apply Please include: 1. Examples of AI or data pipeline projects you have built 2. Experience with LLM-based automation 3. Your proposed architecture for this system 4. Estimated timeline for a first working version