Career-Ops — The AI Job Search System That Landed a Head of Applied AI Role

Santiago Fernandez spent months applying to jobs the hard way after selling his 16-year business. So he engineered the system he wished he had: Career-Ops, a multi-agent job search pipeline built on Claude Code. Paste a job URL, get an A-F evaluation across 10 dimensions, a personalized ATS-optimized PDF resume, and a tracker entry — in minutes. He evaluated 740+ listings, generated 354 tailored CVs, and landed his target role as Head of Applied AI. Then he open-sourced everything.

How It Works

Paste job URL or description
         ↓
Archetype Detection (6 role types)
         ↓
A-F Evaluation (10 weighted dimensions)
         ↓
┌─────────────┬──────────────┬──────────────┐
│ Markdown    │ ATS-Optimized│ Tracker      │
│ Report      │ PDF Resume   │ Entry (TSV)  │
└─────────────┴──────────────┴──────────────┘

Human-in-the-loop by design — it automates analysis, not decisions. It never auto-submits applications. The creator read every report before applying.

Skill Modes

The PDF Engine

Not a template — generates a different resume per job listing:

• Injects JD keywords into your real experience
• Reorders sections by relevance to the specific role
• Typography: Space Grotesk + DM Sans
• Built with Playwright rendering HTML → PDF
• 354 unique PDFs generated during the creator’s search

Production Results

The 10 Evaluation Dimensions

Each job gets scored A-F across:

1. Role Match
2. Skills Alignment
3. Seniority Fit
4. Compensation
5. Company Stage
6. Growth Potential
7. Technical Depth
8. Industry Relevance
9. Location/Remote
10. Culture Signals

The system recommends against applying to anything below 4.0/5 — quality over quantity.

Installation

git clone https://github.com/santifer/career-ops.git
cd career-ops && npm install
npx playwright install chromium
npm run doctor                     # Validates setup
cp config/profile.example.yml config/profile.yml
# Create cv.md with your CV, then:
claude                             # Open Claude Code

How LearnAI Team Could Use This

• Student career services — Integrate into a career prep workshop. Students learn to evaluate job listings systematically (10 dimensions) instead of spray-and-pray.
• CS capstone showcase — The project itself is a masterclass in multi-agent architecture: skills, batch processing, PDF generation, terminal dashboard. Ideal for teaching agent system design.
• Faculty job search — Academic job searches benefit from the same pipeline: evaluate postings, tailor CVs, track applications. The /career-ops deep company research is especially useful for understanding department culture.
• Teaching AI ethics — Discuss the human-in-the-loop design: why the system deliberately doesn’t auto-submit applications. What would go wrong if it did?
• Resume workshops — The ATS optimization approach (keyword injection, section reordering per JD) teaches students what actually matters for getting past automated screening.

Real-World Use Cases

1. Career changers — The archetype detection (6 role types) helps people transitioning between fields understand which roles actually match their experience.
2. Batch job market analysis — Evaluate 100+ listings in parallel to understand market trends, salary ranges, and skill demands in your field.
3. Portfolio proof — The creator used the project itself as proof of competency — building a multi-agent system while searching for multi-agent roles. Meta-level career strategy.
4. Recruitment teams — Reverse the perspective: understand how candidates use AI to optimize applications, and what ATS-gaming looks like from the other side.

Tech Stack