application-pipeline
Career application infrastructure treating job, grant, and residency search as a conversion pipeline. Implements a “Cathedral → Storefront” philosophy: preserving deep, immersive systemic work while providing high-signal, scannable entry points for reviewers.
Motivation
Application processes waste enormous energy on redundant writing. This system treats narrative content as reusable atoms (blocks), composes them into target-specific molecules (submissions), tracks outcomes as conversion data, and uses that data to improve future applications.
Quick Start
pip install -e ".[dev]" # Install with dev dependencies
make verify-quick # Deterministic local verification (python -m pytest)
python scripts/run.py standup # Daily dashboard
python scripts/run.py campaign # Deadline-aware campaign view
python scripts/run.py score <id> # Score a single entry
python scripts/run.py submit <id> # Generate portal-ready checklist
python scripts/run.py --help # All 40+ commands
Pipeline State Machine
stateDiagram-v2
[*] --> research
research --> qualified : score ≥ threshold
qualified --> drafting : blocks assigned
drafting --> staged : content complete
staged --> submitted : portal submission
submitted --> acknowledged : confirmation received
acknowledged --> interview : interview scheduled
interview --> outcome : terminal result
qualified --> deferred : blocked externally
drafting --> deferred
deferred --> qualified : re-activate
deferred --> staged : re-activate
outcome --> [*]
note right of outcome
accepted | rejected
withdrawn | expired
end note
Architecture
pipeline/ One YAML per application (state machine entries)
active/ Actionable entries (research → staged)
submitted/ Post-submission tracking
closed/ Terminal outcomes
research_pool/ Auto-sourced candidates (~940 entries)
blocks/ Modular narrative building blocks (tiered: 60s / 2min / 5min / cathedral)
variants/ A/B tracked material versions (cover letters, project descriptions)
materials/ Resumes (base + tailored batches), CVs, work samples
targets/ Target research, profiles (44 target-specific JSONs)
signals/ Conversion analytics (logs, patterns, outreach)
strategy/ Scoring rubric, identity positions, market intelligence
scripts/ 52 Python CLI tools (all import from pipeline_lib.py)
tests/ 1263 pytest tests
docs/ Architecture rationale and workflow guides
Scoring Model
Every entry is scored across 8 weighted dimensions:
| Dimension | Weight | What it measures |
|---|---|---|
| Mission Alignment | 20% | How well the opportunity matches identity position |
| Evidence Match | 15% | Strength of portfolio evidence for this target |
| Track Record Fit | 15% | Relevant experience alignment |
| Financial Alignment | 10% | Compensation / grant amount viability |
| Effort to Value | 15% | Time investment vs. potential return |
| Strategic Value | 10% | Network effects, learning, positioning |
| Deadline Feasibility | 10% | Can we produce quality work in time? |
| Portal Friction | 5% | Submission complexity and platform friction |
Identity Positions
Every application aligns to one of five canonical positions:
- Independent Engineer — AI lab roles. Focus: infrastructure, testing, AI-conductor methodology
- Systems Artist — Art grants/residencies. Focus: governance as artwork, systemic scale
- Educator — Academic roles. Focus: teaching complex systems at scale
- Creative Technologist — Tech grants/consulting. Focus: AI orchestration, creative instruments
- Community Practitioner — Identity-specific funding. Focus: precarity-informed systemic practice
Session Sequences
| Workflow | Commands |
|---|---|
| Morning | standup → followup → outcomes → campaign |
| Submit | campaign → check <id> → submit <id> → record <id> |
| Research | hygiene → scoreall → qualify → enrichall |
| Analyze | funnel → conversion → velocity → dashboard |
| Strategy | startup → funding → differentiate → tracker |
Development
make verify # Full verification gates (matrix + lint + validate + full tests)
make verify-quick # Faster local verification loop
make preflight # Staged preflight gate
python scripts/verification_matrix.py --strict # Module-to-verification route coverage
python -m ruff check scripts/ tests/ # Lint only
python -m pytest tests/ -v # Full test suite only
python scripts/validate.py # YAML schema validation
CI runs on every push/PR via .github/workflows/quality.yml with a Python 3.11/3.12 matrix; scheduled full regression also uploads verification artifacts.
License
Portfolio · System Directory · ORGAN 4444J99 · Part of the ORGANVM eight-organ system