data-ml-pack
Build trustworthy data and ML systems — reliable pipelines, validated data, leakage-free features, and honest model evaluation, sequenced from raw source to a model you can defend in production.
Included with Vanara Pro ($10/mo) — unlocks all 206 items. See pricing →
Overview
Most ML failures are data failures wearing a model's clothes. The pipeline double-counts, a null sneaks through, a feature quietly leaks the target, the test set overlaps the training set — and a model that looks brilliant offline falls apart the day it serves real traffic. This pack bundles the agents and skills that make each of those failure modes visible *before* they reach production, and — j
What’s inside
One install of data-ml-pack deploys every member below.
Agents: data-pipeline-engineer, data-quality-auditor, ml-model-reviewer
Skills: data-validation, feature-engineering, model-evaluation
What it covers
- Data to deployed model: the discipline playbook
- The six-stage data-to-model workflow