/ Retail Replenishment & Demand Forecasting
A pre-pilot retail inventory decision platform that turns sales, inventory, product, supplier, and integration data into demand forecasts, stockout/overstock risk, replenishment recommendations, and buyer review workflows.
Retail replenishment decisions sit across noisy systems: transactions, inventory levels, supplier rules, receiving issues, purchase orders, promotions, and exception events. A useful system has to connect forecasts to reviewable operational decisions.
Async FastAPI APIs, SQLAlchemy/PostgreSQL tenant context, Redis/Celery workers, integration adapters, demand forecasting artifacts, dataset snapshots, backtests, model registry and promotion gates, simulation endpoints, anomaly workflows, and React/TypeScript decision dashboards.
Python, FastAPI, SQLAlchemy, PostgreSQL, Redis, Celery, React, TypeScript, LightGBM, SHAP, Pydantic, Docker.