Data Science Consultation: AI/ML Strategy for an E-commerce Purchasing Platform

WHO WE ARE We are a Netherlands-based e-commerce business. We buy consumer electronics (smartphones, tablets, smartwatches, laptops) from wholesalers and resell them on Bol.com, the leading online marketplace in the Netherlands and Belgium. We run several companies that sell the same assortment, all managed through one internal system we built ourselves, called GoldSpy. WHAT OUR SYSTEM DOES GoldSpy brings together our market data and our own operational data to answer one question every day: which products are worth buying and reselling, at what margin, and how much can we buy this period. Every day it collects competitor and price data, Bol.com commissions, supplier purchase prices and sales-velocity estimates; it calculates profitability and ROI per product; it supports our daily purchasing decisions; and it tracks our actual sales and returns — including the ROI we actually achieved versus what we expected up front. We have attached a detailed handbook that describes the system, the data it uses, how fresh and reliable that data is, and how we use it day to day. WHAT WE WANT TO DO NEXT (where you come in) We want to start putting our data to work with AI/ML. Our ambition, roughly in stages: - An assistant that answers purchasing questions in natural language (about a product, the market, competitors, or our own history). - Optimizing our purchasing decisions: better predictions of the achievable sell price, sell-through speed and real ROI; flagging opportunities (price drops, temporary commission discounts, a competitor dropping out) and risks (prices that are too volatile to touch); and smarter allocation of our shared budget across our companies. - Moving toward (semi-)automatic purchasing within predefined rules and budgets, with human oversight. WHAT WE ARE LOOKING FOR (this engagement) Right now we are not hiring for the full build yet. We are looking for a paid consultation — a [60-90 minute] call with an experienced data scientist to get honest, practical advice before we commit to a direction. We would like your view on: - Is this realistic, and what would a sensible, phased roadmap look like? - What should we prioritize first (data readiness, quick wins, the assistant, forecasting)? - How ready is our data, and what would you want us to set up or fix before modeling? - What kind of person or team should we hire to build this — profile, seniority, and whether we need a data engineer, an applied data scientist, or both? This is advisory first. If it is a good fit, there may be follow-on work (implementation or ongoing guidance), but the consultation stands on its own and we are happy to pay for your time. ABOUT YOU - Experienced applied / full-stack data scientist — not purely academic research, and not purely BI/reporting. - Strong Python and SQL; comfortable with data engineering and data quality, not only modeling. - Hands-on machine learning with tabular and time-series data: demand and price forecasting, and optimization. - Experience with LLMs / RAG for building assistants on a company's own data. - A strong evaluation mindset — able to show that a model actually improves decisions, not just that it runs. - E-commerce, online marketplace, pricing or retail experience is a big plus. - A clear communicator who can explain trade-offs to a non-technical founder. - Dutch is a plus (our domain is Dutch / Benelux), but not required. TO APPLY Please include in your proposal: - A short note on similar projects you have done — especially e-commerce, pricing or forecasting work, or building data/ML capability from scratch for a small company. - Your initial read on our situation: what stands out to you, and what you would want to dig into. - Your availability for a call [this week / next week], your time zone, and your hourly rate for the consultation.