ML Engineering Workflow
Production ML engineering workflow: data contracts, reproducible training, model evaluation, deployment, monitoring, and rollback.
- What
- Production ML engineering workflow: data contracts, reproducible training, model evaluation, deployment, monitoring, and rollback.
- Cost
- Free
- Needs
- Use "ML Engineering Workflow" with your Muse.
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor: a workflow skill for turning model work into production ML systems. It covers data contracts, reproducible training pipelines, offline and online model evaluation with clear promotion criteria, experiment tracking, deployable artifacts, canary rollouts and shadow traffic, operational monitoring (data drift, label leakage, stale features), and rollback paths when quality degrades. Its scope calibration is a highlight: use only the lanes that fit the system in front of you, whether ranking, recommendations, classifiers, forecasting, embeddings, or LLM workflows, without forcing one architecture onto everything. Debugging guidance targets the classic production failures: data drift, label leakage, artifact mismatch, and training/serving skew. By @affaan-m, listed here with credit to its creator. From the affaan-m/ECC repository (MIT, declared in frontmatter). Honest caveats: pure guidance, nothing to install; assumes ML work beyond one-off notebooks; framework-agnostic, so adapt the lanes to your stack. Skill Harbor never reviews the code, review it yourself before use.
Version:
Install
Copy the install package below, then paste it into MuseThe install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Use "ML Engineering Workflow" with your Muse. Prerequisites: an ML project moving (or already in) production; notebook-only experiments gain little. Pure guidance skill, nothing to install. 1. Open the skill: https://github.com/affaan-m/ECC/blob/main/skills/mle-workflow/SKILL.md and copy the full SKILL.md text. 2. Paste it into a chat with Muse and add: "Review my ML pipeline against this workflow: data contracts, reproducible training, eval gates, monitoring, rollback." 3. For a production incident, ask: "Debug this model degradation using the failure playbook: data drift, label leakage, stale features, or training/serving skew?" Tip: start by naming your serving setup (batch, online, or both); the scope calibration picks the right lanes. Safety: a skill is plain-text instructions; it runs nothing by itself. Never let model changes reach production without the eval gates this workflow describes.
Saved to your recent installs. Find it anytime on /connect.
Questions
How do I install a build?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the build up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a build before installing it.