Skip to content

Notebooks

Interactive tutorials demonstrating Feature Forge capabilities. All notebooks are rendered with Quarto and embedded below.

Notebook Description
Quick Start: Sklearn API Get started with MALMASFeatureEngineer — fit, transform, and integrate into sklearn pipelines.
Agents: The 6 Specialized Feature Generators Explore all 6 built-in agents, the AgentRegistry, and how to build a custom agent.
Router & Pipeline Modes Dynamic agent selection with data-driven, performance-driven, hybrid, and LLM-based router strategies.
Iterative Pipeline & Memory Multi-round feature engineering with procedural, feedback, and conceptual memory tiers.
Experiment Matrix & Tracking Design Cartesian experiments, run them with ExperimentRunner, and report results.
Baselines Comparison Run OpenFE, CAAFE, LLM-FE, and Malmus on the same data and compare artifacts.
Evaluation & Sandboxed Execution Cross-validation feature evaluation, sandboxed code execution, and model factory.
Artifacts & Dashboard Unified artifact access, comparison, diff, and offline HTML dashboards.
Configuration & Tracking Pydantic settings, environment overrides, and experiment tracking with WandB and MLflow.
Method Comparison Head-to-head benchmark of MALMAS, CAAFE, LLM-FE, and Malmus on the same dataset with side-by-side metrics.

Prerequisites

export FF_LLM__API_KEY=sk-your-deepseek-key
uv sync --extra docs