---
name: resume-tailor
description: >-
  Compose a tailored resume for a specific job lead from profile/ inventory.
  Saves to outputs/L0001/. Ryan's voice — concrete projects, no fabrication.
  Use when tailoring a resume, applying to a role, or preparing application
  materials.
---

# Resume tailor

One resume per role. Pull facts from `profile/`, shape for the lead, sound like Ryan.

## Read first

1. **`profile/voice.md`** — mandatory (projects are proof points, not identity)
2. Target lead: `leads/L0001-*.md` + `companies/CO0001-*.md`
3. `profile/experience.md`, `profile/projects.md`, `profile/education.md`, `profile/leadership.md`, `profile/skills.md`
4. **`goal.md`** — what kind of work this company does; pick projects that connect to *their* problems

Optional refresh: [ryanhub.org/me](https://ryanhub.org/me), [resume.pdf](https://ryanhub.org/resume.pdf)

## Output

Save to:

```
outputs/L0001/resume_notes.md
```

Update lead frontmatter:

```yaml
outputs:
  resume: outputs/L0001/resume_notes.md
status: drafted
```

Then sync:

```bash
.venv/bin/python scripts/sync_indexes.py
```

When Ryan actually submits:

```bash
.venv/bin/python scripts/log_event.py \
  --type application --lead L0001 \
  --summary "Submitted via company portal"
```

## Structure

```markdown
# Ryan Alport

[email protected] · 908-335-4755
ryanhub.org/me · linkedin.com/in/ryanalport · github.com/RyanA005

## Summary

2–3 sentences. Ryan's voice. Tie to **this company's technical work** — ambitious builder/researcher angle, not a fixed project pitch.

## Experience

Pick relevant roles from profile. 2–4 bullets each. Metrics where true.

## Projects

Pick 2–4 projects that match **this posting's domain**. Vary selections across applications — do not default to the same flagship every time.

## Skills

Subset from profile/skills.md — match job description, no keyword dump.

## Education

NJIT, May 2027, GPA 3.8, minor Applied Physics. Coursework only if relevant.
```

## Selection rules (pick by company type — do not repeat same lead every time)

- **AI research / product lab:** baby-lm, baby-image-gen, assistant agent, CNAlab, Fivebase cost work — emphasize research-engineering, prototypes, eval mindset
- **Devtools / agents / infra:** api-frame, RyanHub, tooling/harness work, agent-related experiments — emphasize systems that make builders more effective
- **Robotics / simulation / embodied AI:** physics/simulation projects, vision/control work, game engines, spatial reasoning — only if in profile and relevant
- **Systems / low-level:** y86 IDE, couch-coop, cell phone project, simple-game, from-scratch C work
- **Backend / infra:** api-frame, Fivebase, RyanHub, Flask/backend work
- **Startup / high-agency generalist:** strongest ownership bullets from whatever domain matches *their* product — founding experience OK if relevant, not automatic lead

Drop irrelevant projects — one page target for interns. **Avoid** leading with Ezcaptain/maritime unless the company is specifically in vertical SaaS or ops tooling.

## Voice

- "Built X from scratch in C" beats "Experienced with C"
- Name projects when relevant — not "personal project"
- Summary should not read like a project catalog
- No fabricated bullets. If gap exists, address honestly or omit

## Do not

- Copy one generic resume for every role
- Use LLM buzzwords Ryan wouldn't say
- Include every project from profile/
- Mark applied without Ryan confirming (use `log_event.py`)