Palantir Foundry · case study

AI-Powered Inventory Demand Forecaster

Six-month, per-SKU demand forecasts trained in Foundry Model Studio, with approved signals flowing straight into the production planning channel.

← All work

What I built

  • —Joined SKU master, historical and open orders, quotes, sales and cost metrics with an external oil-price index into one model-ready dataset.
  • —Partitioned SKUs into regular, intermittent and sparse demand profiles (ADI / CV²) and trained a dedicated forecasting pipeline for each.
  • —Validation and production forecast pipelines feed Workshop dashboards with predicted min/max levels.