Description
Summary Senior Full-Stack / AI Developer Needed – E-commerce Product Automation System Project Overview We are a Scandinavian e-commerce company operating in Sweden, Norway, Denmark and Finland with a large product catalogue and new products being added almost every day. We are looking for an experienced Full-Stack Developer / AI Automation Engineer to build an internal system that significantly automates our current product creation workflow. Our team currently manages new products through Google Sheets, which is then imported into our e-commerce platform. We want to keep our existing Google Sheets/import workflow, but build an intelligent system around it that automates the most time-consuming parts of the process. The main areas we want to automate are: * SEO keyword research * Product title generation * SEO product description generation * Product categorization * Competitor product matching * Competitor price research * Recommended selling prices * Human review and approval The goal is not to completely replace human review. The goal is to let AI and automation perform most of the repetitive work while our product writers review, edit and approve the final result. ⸻ Our Current Workflow When we receive a new product from a supplier, our team currently: 1. Adds the supplier/product information to Google Sheets. 2. Researches suitable Swedish product names. 3. Performs keyword research to understand what customers search for. 4. Writes an SEO-optimized Swedish product description. 5. Adds product specifications such as dimensions, weight, material, quantity, etc. 6. Manually selects relevant categories from our existing category tree. 7. Adds the purchase price. 8. Our existing pricing formula calculates a suggested selling price. 9. Our staff manually searches competitors to find the same/similar product and compare prices. 10. The final information is entered into Google Sheets. 11. The products are imported from Google Sheets into our e-commerce platform. We want to automate as much of steps 2–9 as reasonably possible. ⸻ What We Want to Build 1. Google Sheets Integration The system should read new products from our existing Google Sheet. Available information may include: * SKU * EAN/GTIN * Supplier * Supplier product name * Supplier description * Purchase price * Dimensions * Weight * Material * Image URL * Supplier SKU * Other product attributes After processing and human approval, the system should write the completed information back into the correct Google Sheet columns. Our existing product import can then continue working as it does today. ⸻ 2. AI Product Title Generation The system should generate optimized Swedish product titles based on: * Supplier information * Product attributes * SEO/search data * Our existing naming conventions * Examples from our current catalogue Our staff should be able to: Approve / Edit / Regenerate the suggested title. ⸻ 3. SEO & Keyword Research We want the system to help identify relevant Swedish search terms for each product. Possible data sources could include: * Google Search Console * Google search data/results * Existing products on our website * External keyword data/services * AI/LLM analysis We are open to recommendations from the developer regarding the best technical solution. The selected keywords should then be naturally incorporated into the generated product title and description. ⸻ 4. AI Product Description Generation The system should generate Swedish product descriptions according to our existing writing style and structure. We already have: * Writing guidelines * Existing product examples * Required structure * SEO requirements * Product information format * Category structure We can provide examples of high-quality existing product pages. The AI should follow these rules consistently rather than generating generic AI product descriptions. ⸻ 5. Product Specifications The system should structure available product information such as: * Dimensions * Weight * Material * Quantity * Packaging information * Other relevant attributes Missing information should be clearly identified rather than invented by AI. ⸻ 6. AI Product Categorization We already have an established category tree. The system should analyze each product and recommend the most relevant existing categories. Important: AI should only select categories that actually exist in our category database/tree. It should not create new categories. Multiple categories may be relevant to one product. Our staff should be able to approve, remove or add categories before final approval. ⸻ 7. Competitor Product Matching & Price Research This is one of the most important parts of the project. Today our staff manually searches competitors to determine whether we are priced too high or too low. We want the system to investigate whether this process can be automated. The system should attempt to identify the same product at selected competitors using available identifiers and matching methods such as: * EAN/GTIN * SKU / manufacturer number * Product title * Supplier/manufacturer information * Product attributes * Image similarity where useful * AI/fuzzy product matching We will provide a list of relevant competitors. For example, the system could display: Our calculated price: 129 SEK Competitor A: 119 SEK Competitor B: 129 SEK Competitor C: 109 SEK Recommended price: 119 SEK We understand that competitor websites differ technically and that scraping may not always be possible or appropriate. We therefore expect the developer to recommend a reliable and maintainable architecture using APIs/data sources where available and compliant collection methods where appropriate. ⸻ 8. Product Matching Confidence We would like the system to show how confident it is that a competitor product is actually the same product. For example: EAN exact match – 100% or AI/fuzzy match – 72% – Manual review recommended This is important because we do not want incorrect competitor products affecting pricing decisions. ⸻ 9. Pricing Recommendation We already have an internal pricing formula based on purchase price. The system should combine our calculated price with available competitor pricing information and provide a recommended selling price. The final decision should initially remain with our staff. The system should not automatically publish or change prices without approval. ⸻ 10. Internal Review Dashboard We want a simple and efficient internal interface for our product writers. For each new product, the employee should be able to see: * Product image * Supplier information * AI-generated product title * SEO keywords * AI-generated description * Product specifications * Suggested categories * Purchase price * Our calculated selling price * Competitor prices * Competitor matching confidence * Recommended selling price Each section should allow actions such as: Approve / Edit / Regenerate Finally, the employee should be able to click something similar to: APPROVE PRODUCT The approved data should then be written back to Google Sheets. ⸻ 11. Learning From Human Corrections We want to store both: AI suggestion → Human-approved final version For example, if AI recommends one product title but our employee changes it, both versions should be stored. The same applies to: * Product titles * Descriptions * Categories * Keywords * Pricing decisions The purpose is to analyze these corrections and continuously improve the system’s prompts, rules and recommendations over time. We do not necessarily require custom model training in the first version. We are interested in a practical architecture that allows the system to improve based on historical corrections. ⸻ Technical Experience We Are Looking For You should ideally have strong experience with several of the following: * Python * Full-stack web development * LLM / AI API integrations * Google Sheets API * Google Search Console API * REST APIs * Databases * Web data extraction / scraping * Product matching * Fuzzy matching / embeddings * E-commerce systems * AI prompt architecture * Internal dashboards/admin systems Experience building e-commerce automation, PIM systems, product enrichment systems or AI content workflows is highly valuable. ⸻ Important: We Are Looking for a Long-Term Developer This project will likely be developed in stages. We want to start with a solid MVP and then continue improving and expanding the system. There may be additional automation projects in the future if the cooperation works well. We therefore prefer someone who is interested in a long-term working relationship, not only completing a quick one-time task. ⸻ When Applying Please answer the following questions: 1. Have you previously built an AI-powered product management, PIM, e-commerce automation or content generation system? If yes, please describe it and provide examples/screenshots if possible. 2. How would you technically approach competitor product matching? Especially when EAN is unavailable. 3. How would you approach automated competitor price collection while keeping the solution reliable and maintainable? 4. What AI/LLM architecture would you recommend for product title, description and category generation? 5. Have you previously worked with Google Sheets API and Google Search Console API? 6. Would you recommend building this as a custom web application/dashboard, or would you suggest another architecture? Why? 7. What would you include in the first MVP? Please briefly describe your proposed architecture, estimated development time and estimated cost. ⸻ About Us We are an established Scandinavian e-commerce company with a large product catalogue and new products being added continuously. Our objective is to use AI and automation to significantly reduce manual product administration while maintaining human quality control. If the first project is successful, there is potential for substantial additional development and automation work.