Multi-Model ML Framework
30+ parameter framework for runtime algorithm selection, feature group configuration, and model stacking across the forecasting platform.
Akash Sharma
Primary Owner
30+
Config Parameters
4
Algorithm Options
The Problem
Each customer had different data volumes, feature availability, and modeling requirements. A hard-coded single-algorithm approach couldn't serve diverse enterprise customer configurations.
What Was Built
Built a configurable framework with 30+ parameters controlling algorithm selection (XGBoost, Ridge, LightGBM, CatBoost), feature group inclusion/exclusion, ensemble composition, validation split strategy, and scoring behavior. Runtime config parsed from JSON/base64-encoded arguments, enabling per-customer model configuration without code changes.
Business Impact
Eliminated code duplication across customer configurations, enabled A/B testing of model variants, and supported diverse enterprise customer needs from a single codebase.
Related Revenue Forecasting Projects
Revenue Forecasting Platform
Production ML system predicting pipeline and booking outcomes at daily frequency across current and future quarters.
Time-Decay Forecast Smoothing
Dynamic adjustment layer that re-weights prior EOQ history by recency, capturing momentum and velocity signals closer to quarter end.
Seasonal Index Forecast Adjustment
Statistical fallback layer using day-of-quarter index averages when ML model signals are insufficient.