Job Description We need someone to build ML-based forecasting and inventory optimization systems for our Amazon-focused retail/eCommerce business. We have 100+ brands, thousands of SKUs, and currently rely on manual purchasing decisions that are causing stockouts and overstock issues. You'll work with our existing development team (2 dedicated developers) and our existing purchasing team. You'll have access to years of clean historical data on numerous data points. The goal: Build systems that help automate and allow us to make better, more informed purchasing decisions and inventory management decisions. Rough project scope: What You'll Build (In Priority Order) Month 1: • Demand forecasting model that accounts for seasonality, stockouts, and competitor behavior • Safety stock calculator based on lead times and forecast uncertainty • Daily alert system for stockout risks Month 2: • Buy box share prediction model (we compete with other sellers) • Competitor inventory monitoring system • Opportunity scanner for strategic purchases Month 3: • Price elasticity models by product category that apply to inventory management/forecasting • Automated purchasing decision engine • Full integration with existing systems You will need to build documentation on how to use any new tools as well to help our existing team (junior analysts) implement them into their workflow. You Have Access To: • 2 full-time developers for implementation support • 2+ years of historical sales data (already clean and accessible) • Current inventory management team for domain knowledge • Direct access to decision makers (no bureaucracy) • Our existing PowerBI and internal software systems Technical Requirements: Must Have: • Experience with inventory/supply chain optimization • Ability to build production-ready models, not just prototypes • Amazon, retail, or eCommerce marketplace experience (you understand how inventory flows through the supply chain) • Knowledge of competitive dynamics in marketplaces • Experience with automated decision systems • To Note: Before full engagement, we'll pay you for a 4-hour test project: • We'll provide 6 months of sales data for 20 SKUs • Build a basic forecasting model • Identify 3 strategic insights from the data • Present your approach and findings Working Arrangement: • Remote, but available for CST time zone meetings • Start within 2 weeks • 20-40 hours/week depending on your availability • Weekly sync meetings, otherwise autonomous • Delivery-focused (we care about working models, not reports) • What We're NOT Looking For: • Consultants who only make recommendations • People who need perfect data to begin • Anyone who can't show actual models they've built • Theory without implementation ability You're the person who: • Thinks about optimal betting strategies in poker or daily fantasy sports • Notices patterns in everyday life that others miss • Gets genuinely excited when you find an inefficiency to exploit • Has strong opinions about why most people play Monopoly wrong • Sees business as a multiplayer strategy game This job is perfect if you enjoy figuring out how to beat your friends at strategy games, find yourself calculating expected value in daily decisions, or have ever built a spreadsheet to optimize something in your personal life "just for fun." You probably read about game theory for enjoyment, have opinions about prediction markets, and immediately understand why buying a competitor's allocation during supply shortages is brilliant. The Freedom You'll Have: • No committees. No daily standups. You'll have direct access to the CEO, clean data, 2 developers to implement your ideas, and the autonomy to build solutions your way. I care about results, not process. If your model prevents stockouts, reduces inefficiencies, and improves margins, I don't care if you built it at 2 AM in your pajamas. • We're looking for someone who can deliver results quickly. Rate is negotiable for the right expertise. This could convert to a full-time senior role for the right person.