back to projects
AI / Data

g.cast

AI-assisted forecasting for production and pricing.

PythonTime-series MLFastAPISupabase
demo coming soonin-progress
overview

A forecasting layer that sits on top of the production and price data warehouse. Combines classical decline curves, gradient-boosted residual models, and macro features (rig count, frac spread, oil futures curve) to produce probabilistic forecasts at the well, area, and basin level.

highlights
  • Probabilistic forecasts (P10 / P50 / P90) on every series
  • Backtest harness with rolling-origin evaluation
  • Feature store keyed on API number and grid cell
  • Drift monitoring on incoming data