Machine learning model development

Assign model work and the engineer builds a reproducible pipeline with baselines, leakage checks, metric-driven evaluation, error analysis, and a review-ready PR or report.

Saves you ~4 h / run

How it works

Trigger
When you start the “Machine learning model delivery” workflow.
Job
Build and verify a reproducible model result.
Outcome
A tested implementation and pull request draft.

What it installs

Agents 2

  • ML Engineer

    Defines baselines, guards against leakage, evaluates models, writes error analysis, and drafts a PR or report for review.

  • Machine learning model development Quality Reviewer

    Checks primary evidence, domain controls, deliverable completeness, and communication quality, stopping the run when the work is wrong, unsupported, incomplete, or uncertain.

Teams 1

  • Machine learning model development quality team

    The delivery agents produce the work while an independent quality reviewer checks each workflow handoff against explicit evidence, domain, and communication requirements before the run can continue.

Workflows 1

  • Machine learning model delivery

    Build and independently verify a reproducible model result before human review.

Goals 1

  • Models evaluated before shipping

    Keep model changes reproducible, baseline-backed, and honestly evaluated. Success looks like: Every model task has a baseline, leakage-safe pipeline, declared metric, held-out evaluation, error analysis, and a reviewable document before any deployment decision.

Skills 1

  • scikit-learn

    Classical ML workflow guidance for scikit-learn pipelines, preprocessing, model selection, cross-validation, metrics, leakage prevention, and reproducible experiments. Adapted from jackspace/claudeskillz/scikit-learn.

Requirements

  • Connected repository or notebook workspace — Needs access to the modeling code, notebooks, experiment scripts, or pipeline repository where the change should land.
  • Dataset and target metric — Needs a dataset or sample, the prediction target, and the metric that decides whether the model is better.

Setup guide

How to Develop Machine Learning Models

A practical guide to model development with baselines, leakage-safe pipelines, metric-driven evaluation, error analysis, and review.

Read the setup guide

Don't see your workflow? Describe it.

A sentence or two about a recurring job is enough. We design the playbook that runs it and show you exactly what it saves.

* What keeps taking time you don't have? *



Takes a minute · no account needed