b02-skills-main-datascience

KO | EN

gitlite — search

#python #javascript #telegram-bot #library #android #chat #bot #language #translation #chrome #extension #github

b02-skills-main-datascience

★ 48 Open GitHub ↗

🤖 Data Science Engineering Skills — TDD and planning skills for ML pipelines, data APIs and analytical tooling.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# 🤖 Data Science Engineering Skills

![Domain](https://img.shields.io/badge/Domain-Data%20Science%20&%20AI/ML-purple?style=for-the-badge)
![Skills](https://img.shields.io/badge/Skills-8-blue?style=for-the-badge)
![Workflows](https://img.shields.io/badge/Workflows-3-orange?style=for-the-badge)
![Source](https://img.shields.io/badge/Source-skills-main%20(mattpocock/skills)-grey?style=for-the-badge)
![License](https://img.shields.io/badge/License-MIT-green?style=for-the-badge)

> **TDD and planning skills for ML pipelines, data APIs and analytical tooling**

Adapted from **skills-main (mattpocock/skills)** — _real-engineering TDD and planning skills for professional developers_

---

## What This Does

Data pipelines, model training, evaluation, MLOps.

This collection brings the **Data Science Engineering Skills** approach to Data Science & AI/ML work,
providing **8 domain-specific skills** and **3 end-to-end workflows**
with structured output and live progress tracking.

---

## Skills

| Skill | Description |
|-------|-------------|
| `/pipeline-tdd` | TDD cycle for ETL pipelines — test data contracts and transformations first |
| `/model-prd` | Turn ML experiment findings into a structured model deployment PRD |
| `/ds-to-issues` | Break data platform work into independently-testable GitHub issues |
| `/feature-design` | Design multiple feature engineering approaches via parallel agent analysis |
| `/etl-refactor` | ETL refactor plan with data-contract compatibility and migration steps |
| `/hypothesis-grill` | Interview mode for ML hypothesis validation and experimental design |
| `/data-triage` | Triage data quality issues through label state machine with dbt test links |
| `/eval-tdd` | Model evaluation TDD — define metric thresholds as tests before training |

---

## Workflows

| Workflow | Description |
|----------|-------------|
| `ml-sprint` | ML feature sprint: hypothesis → PRD → data issues → TDD → eval → deploy |
| `pipeline-rebuild` | Data pipeline rebuild: contract design → vertical slices → TDD → monitor |
| `model-launch` | Model launch: eval design → shadow mode → canary → full rollout → monitor |

---

## UI Preview

```
╔══════════════════════════════════════════════════╗
║  Pipeline TDD  —  "user event aggregator"        ║
╠══════════════════════════════════════════════════╣
║  🔴 RED    Test: null events don't break schema  ║
║  🟢 GREEN  Nullable handling added               ║
║  🔵 REFACTOR Extract EventValidator class        ║
╚══════════════════════════════════════════════════╝

Slice #3/6: "deduplicates events within 5s window"
  Data contract: input schema + idempotency key defined
  Status: ✓ GREEN — 4 tests, real Kafka consumer stub
  Test coverage: 87% on core transformation logic
```

---

## How It Works

Every skill follows the same 5-step interaction:

```
① Scope Confirmation   verify target + options
② Live Progress        [████████░░] 80%
③ Structured Findings  table sorted by impact (🔴🟠🟡🟢)
④ Action Plan          quick wins → medium-term → strategic
⑤ Next Steps           suggested follow-up skills
```

---

## Install

```bash
# Copy to Claude skills directory
cp -r . ~/.claude/skills/b02-skills-main--datascience/

# Load in Claude Code session
/read ~/.claude/skills/b02-skills-main--datascience/SKILL.md
```

## Source

Derived from **skills-main (mattpocock/skills)** — real-engineering TDD and planning skills for professional developers.
Original patterns adapted and domain-specialised for Data Science & AI/ML.

---
MIT License

b02-skills-main-datascience

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README