AI Data Specialist

Upwork Post — Product Data & AI Specialist We are an interior design and home furnishings company looking for a highly detail-oriented Product Data & AI Specialist to clean and standardize furniture vendor CSV catalogues for our e-commerce website. You will be provided with a detailed training guide and established workflow. The role requires using AI, Excel formulas, scripts and batch processing to efficiently standardize product names, descriptions, categories, materials, dimensions, pricing, SKUs and other product data, followed by careful human review, exception handling and final quality control. This is not manual product-by-product data entry. We are looking for someone who understands how to use AI and automation to process large datasets efficiently while still taking personal responsibility for the accuracy and quality of the finished files. We currently have 3 vendor catalogues ready to begin and approximately 20 additional vendors to follow, so there is significant potential for ongoing work. Compensation This is fixed-price project work, not hourly or per-SKU work. Our typical vendor catalogue contains approximately 1,000 products, with an established budget of approximately $200 per completed vendor catalogue. Catalogue sizes vary. Some may contain 800 products, while others may contain 2,000–2,500+. Because our workflow relies heavily on AI, formulas, scripts and batch processing, compensation does not increase proportionally with SKU count. A catalogue containing 2,000 products, for example, should not require twice the human work of a 1,000-product catalogue. As a general guideline: Catalogue Approximate Budget ~800 relatively straightforward SKUs $150–$175 ~1,000 typical SKUs ~$200 ~2,500 SKUs ~$250–$325, depending on complexity and additional human work Larger or more complex catalogues may warrant a higher fixed price based on the actual additional human work required for setup, exceptions, verification and quality control, rather than simply the number of SKUs. Expected Workflow The workflow is designed so that AI and automation handle the bulk repetitive work, while human time is focused on setup, judgment, exceptions, verification and final QA. For a typical catalogue of approximately 1,000 products, the workflow should look approximately like this: Work Approx. Human Time Inspect raw scrape, understand columns & set up master file 30–60 min Set fixed fields, formulas & transformations 30–60 min Run AI category/material/dimension batches 1–2 hrs Product-family grouping & IG naming 30–60 min Run AI descriptions & specifications 1–2 hrs Pricing & SKU formulas 15–30 min Generate/format final deliverables 30–60 min Human review, exceptions & corrections 3–5 hrs Final QA/checklist 1–2 hrs Typical total human working time ~8–15 hrs These times are guidelines rather than strict limits. Accuracy and production-ready quality are essential, but we are also looking for someone who can develop an efficient, scalable workflow rather than manually processing hundreds or thousands of products individually.