| name | freelancer-project-delivery |
| description | Use when onboarding clients, setting up git repositories, designing MLOps pipelines (ZenML, MLflow, DVC), structuring tests/QA for ML, and managing project milestones or delivery handoffs. |
Freelancer Project Delivery System
Deliver production-grade work that exceeds client expectations, maintains quality standards, and creates natural upsell opportunities through systematic project management and MLOps excellence.
Skill Architecture
To utilize this skill's core workflows, configuration patterns, or checklists, read the corresponding document in the references/ directory:
| File | Focus | Purpose |
|---|
| 04A-KICKOFF-SYSTEM.md | Kickoff System | Client onboarding checklist, technical access protocol, communication bounds, project portal configuration |
| 04B-PROJECT-MANAGEMENT.md | Project Management | Agile sprint structures, milestone definitions, async weekly status templates |
| 04C-DEVELOPMENT-WORKFLOW.md | Development Workflow | Git branch architecture, pull request templates, linter and formatter settings |
| 04D-ML-OPS-PIPELINE.md | MLOps Pipeline | Production ML pipelines (ZenML, MLflow, DVC), model registration, drift detection (Evidently AI) |
| 04E-TESTING-QA.md | Testing & QA | unit testing, integration tests, robustness testing, CI pipelines |
| 04F-DELIVERY-HANDOFF.md | Delivery & Handoff | Production handoff checklists, demo video creation guide, live training scripts |
Agent Guidelines & Workflows
1. Onboarding & Kickoff
When initiating a new client contract:
- Reference 04A-KICKOFF-SYSTEM.md to set up the client's Project Portal (Slack/Notion/Linear) and request infrastructure/data access.
- Execute the kickoff checklist to define communication boundaries (no unscheduled calls, weekly async reports).
2. Agile Project Management
When managing timelines and writing client updates:
- Reference 04B-PROJECT-MANAGEMENT.md to run 2-week development sprints featuring clean milestone tracking.
- Draft professional weekly status updates detailing what was done, what is planned next, and blockages/risks.
3. Engineering & Development Workflows
When setting up codebases or code review pipelines:
- Reference 04C-DEVELOPMENT-WORKFLOW.md to establish clean git workflows (
main $\rightarrow$ dev $\rightarrow$ feature/*).
- Configure automated code formatting (Black, Ruff) and apply standardized pull request templates.
4. MLOps System Design
When designing, building, or configuring machine learning pipelines:
- Reference 04D-ML-OPS-PIPELINE.md to structure model runs with ZenML, version data with DVC, track parameters/metrics with MLflow, and registry/deploy models.
- Set up automated data and concept drift detection using Evidently AI.
5. ML-Specific Testing and Quality Assurance
When writing testing pipelines for data or models:
- Reference 04E-TESTING-QA.md to establish a test suite containing: python unit tests, data schema validation (Pydantic), model performance assertion checks, and robustness/bias verification.
6. Client Handoff and Demos
When completing the project and transferring ownership:
- Reference 04F-DELIVERY-HANDOFF.md to record high-impact 2-minute demo videos, package documentation, and run live training sessions.
Quick Reference Summary
| Concept | File | Section | One-Line Summary |
|---|
| Project Portal | 04A | 3 | Establish a shared workspace as the single source of truth for communication |
| 2-Week Sprints | 04B | 2 | Structure projects into bi-weekly sprints with concrete, demoable deliverables |
| Git Branch Flow | 04C | 2 | Protect production branch: branch from dev, require PR verification and linting |
| Production ML Pipeline | 04D | 2 | Maintain reproducible pipelines: Data Validation $\rightarrow$ Feature Prep $\rightarrow$ Model Registry |
| Drift Dashboards | 04D | 5 | Implement data drift alerts to flag performance decay in production |
| Testing Pyramid | 04E | 2 | Structure test suite: 70% Unit tests, 20% Integration tests, 10% End-to-End checks |
| 2-Minute Demo | 04F | 2 | Record a concise, problem-to-solution video demo of the working system |
Previous Skill: freelancer-sales-closing ←
Next Skill: freelancer-post-delivery →
