Skip to content

Firefly DataScience

Firefly DataScience — AutoML that fuses GenAI with classical ML & Deep Learning

Quick Start Tutorial Configure the LLM

AutoML that fuses GenAI with classical ML & Deep Learning — hexagonal, secure-by-default, and native to the Firefly Framework.

fireflyframework-datascience is a state-of-the-art Python metaframework for AutoML. It pairs GenAI — built on fireflyframework-agentic, which wraps Pydantic AI — with traditional ML and Deep Learning, so any team can apply data science to any project quickly, with production governance, hexagonal swappability, and security by default.

The reproducible pattern — the LLM proposes; the classical engine decides

GenAI proposes code, features, pipelines and seeds; a deterministic classical engine trains, scores and selects; and every GenAI step is gated behind a measured improvement over a seeded classical baseline. GenAI is a governed, measurably-gated accelerator over a battle-tested classical core — never a black box.

Firefly DataScience architecture: GenAI proposes, a classical engine decides, behind hexagonal ports

Want the whole story in one document?

The Complete Guide (PDF) combines the executive summary and strategic case (faster time-to-value, governed GenAI, no lock-in) with the full architecture, a hands-on tutorial, and the benchmark evidence — one document for both leaders and engineers.

Why Firefly DataScience?

  • Classical-first AutoML

    A deterministic engine trains, scores and selects across scikit-learn, XGBoost, LightGBM, CatBoost, AutoGluon and TabPFN — reproducible from a seed.

    Classical AutoML

  • GenAI as a gated accelerator

    The LLM proposes features and pipelines; nothing ships unless it beats the seeded classical baseline (genai.cost_benefit_gate is on by default).

    GenAI features

  • The agentic ML-engineering loop

    Propose → train → score → select, driven by the agentic runtime, with a measured improvement required at every step.

    Agentic loop

  • Deep Learning, swappable

    PyTorch Lightning and HuggingFace sit behind the same ports as the classical adapters — tabular, text, vision, timeseries and multimodal.

    Deep Learning

  • Hexagonal & swappable

    Every ML/MLOps library (MLflow, Feast, BentoML, …) is a swappable adapter behind a Protocol port; the core stays library-agnostic.

    Architecture

  • Secure by default

    LLM-generated code runs in a sandbox (monty by default) with timeouts and approval gates; GenAI is off until you enable it.

    Security model

  • Explainable & trustworthy

    Deterministic global + local feature importances (permutation, SHAP) and calibrated probabilities — so every model can be explained, and its scores trusted for real decisions.

    Explainability

Get started in 30 seconds

uv add fireflyframework-datascience                    # core (ports, app, DI — no heavy ML libs)
uv add 'fireflyframework-datascience[automl-stack]'    # + classical AutoML + tracking

Requires Python 3.13+. Extras compose, e.g. [tabular,tracking,genai].

from fireflyframework_datascience import FireflyDataScienceApplication

# load config -> print banner -> wire DI container -> wiring summary -> ready context
app = FireflyDataScienceApplication.run()

print(app.bean_count)                    # number of wired beans
print(app.config.default_ml_framework)   # "sklearn"
print(app.applied_auto_configurations)   # discovered auto-configurations

firefly-ds doctor       # check your environment & installed adapters
firefly-ds introspect   # boot the app and show discovered auto-configurations

Full quick start

GenAI is classical-first and off by default — opt in, and require a measured win, explicitly:

config = FireflyDataScienceConfig(app_name="lumen-credit-risk", default_ml_framework="lightgbm")
config.genai.enabled = True              # opt in to the GenAI accelerator
config.genai.cost_benefit_gate = True    # require a measured win over baseline
config.execution.sandbox = "docker"      # sandbox LLM-generated code

app = FireflyDataScienceApplication.run(config=config)

Explore the docs

  • Architecture

    Hexagonal ports/adapters, the DI container, and auto-configuration.

  • Quick Start

    Install, boot an ApplicationContext, run your first AutoML job.

  • Configuration

    FireflyDataScienceConfig, profiles, env vars, YAML overlays.

  • Datasets

    Dataset backends (pandas, …) and Modality.

  • Classical AutoML

    Train, score, select — with calibration, stacking ensembles, PR-AUC selection & CV strategies.

  • Explainability

    Deterministic global + local feature importances (permutation, SHAP).

  • GenAI features

    The gated GenAI accelerator and the cost-benefit gate.

  • Agentic loop

    Propose → train → score → select on the agentic runtime.

  • Deep Learning

    PyTorch Lightning & HuggingFace behind the ports.

  • Serving

    Model registry, feature store, and BentoML serving.

  • Security

    Sandboxed code execution, approval gates, secure defaults.

  • Benchmarks

    Reproducible measurement of GenAI vs. classical baselines.

  • Use case: Lumen

    End-to-end lending vertical worked example.