How to Customize Archetypes and Scoring Weights for Your Career Field in Career-Ops

To customize archetypes and scoring weights in Career-Ops, edit the modes/_profile.md file to define your target roles and override default weights, while keeping config/profile.yml synchronized with your target role list.

Career-Ops evaluates job offers through a layered scoring system that separates system defaults from user-specific customizations. The archetype table defines the roles you target, while scoring weights control how each evaluation signal contributes to your final score. Both configurations live in modes/_profile.md, ensuring your personal career preferences stay synchronized across every evaluation cycle.

Understanding the Evaluation Architecture

Career-Ops implements a three-tier configuration system. The base layer resides in modes/oferta.md, which defines a 10-dimension scoring matrix with default weights of 1.0 for dimensions like Compensation, Growth Potential, and Technical Fit.

The middle layer, modes/_shared.md, contains global defaults applied across all evaluations.

Finally, modes/_profile.md serves as the user-specific override layer. According to the pipeline rules documented in _shared.md, the system reads cv.md, then config/profile.yml, and finally modes/_profile.md to apply your personal customizations last.

Customizing Archetypes in modes/_profile.md

Archetypes represent the specific roles you want to evaluate, such as "Data Engineer" or "ML Engineer". To define these, edit the markdown table in modes/_profile.md as documented in docs/CUSTOMIZATION.md.

The table requires three columns:

• Archetype: The role title
• Thematic axes: Key skills or focus areas (e.g., "data pipelines, ETL")
• What they buy: The value proposition that matters to this role

| Archetype | Thematic axes | What they buy |
|-----------|---------------|---------------|
| **Data Engineer** | data pipelines, ETL, cloud storage | Reliable, scalable data infra |
| **ML Engineer** | model training, feature pipelines | Production-ready ML systems |
| **Product Manager** | roadmap, UX, metrics | User impact and market fit |

Overriding Scoring Weights

After the archetype table, add a "Scoring Weights" section to override the defaults from modes/oferta.md. Each dimension listed here receives a numeric multiplier that scales its contribution to the final score.

Available dimensions include:

• Compensation
• Growth Potential
• Team & Culture
• Technical Fit
• Impact Scope

## Scoring Weights

| Dimension | Weight |
|-----------|--------|
| Compensation | 1.5 |
| Growth Potential | 1.2 |
| Team & Culture | 1.0 |
| Technical Fit | 2.0 |
| Impact Scope | 1.3 |

When present, these values replace the built-in weights defined in modes/oferta.md. The evaluator multiplies each dimension's raw score by your specified weight before calculating the final total.

Synchronizing Target Roles in config/profile.yml

To ensure the scanner only pulls relevant opportunities, update config/profile.yml with your target roles under the target_roles key. This list must align with the archetypes defined in modes/_profile.md.

target_roles:
  - Data Engineer
  - ML Engineer
  - Product Manager

Keeping these files synchronized ensures the pipeline filters incoming offers against your north-star roles before running the weighted evaluation.

Runtime Execution Flow

During evaluation, Career-Ops processes your configuration in sequence:

1. Loads your CV data from cv.md
2. Reads personal settings from config/profile.yml
3. Applies custom archetypes and weights from modes/_profile.md (overriding defaults)
4. The evaluator in modes/oferta.md builds the 10-dimension scoring matrix, applying your custom multipliers
5. Generates the final report (blocks A–F) displaying weighted scores
6. Determines whether to auto-generate a PDF or flag the offer for follow-up based on score thresholds

When to Edit System Defaults

User-specific tweaks belong exclusively in modes/_profile.md. If you need to change global settings for every user—such as adding an 11th scoring dimension or modifying the base prompt—edit modes/_shared.md and batch/batch-prompt.md instead.

Never place personal archetype definitions or individual weight preferences in the shared files, as these changes would affect all users of the system.

Summary

• Define archetypes in modes/_profile.md using a markdown table with thematic axes and value propositions
• Override scoring weights in the same file's "Scoring Weights" section to prioritize dimensions critical to your field
• Sync target roles in config/profile.yml to ensure the scanner filters for relevant opportunities
• Reserve system edits for modes/_shared.md and batch/batch-prompt.md only when changing global defaults
• Runtime application: The pipeline automatically reads your customizations after defaults, applying your weights to the 10-dimension matrix from modes/oferta.md

Frequently Asked Questions

Can I add new scoring dimensions beyond the default 10?

Yes, but this requires editing system-wide files. Add new dimensions to modes/_shared.md and update batch/batch-prompt.md to include the new criteria in the evaluation prompt. User-specific weight overrides in modes/_profile.md can only modify existing dimensions defined in modes/oferta.md.

What happens if I don't specify custom weights in modes/_profile.md?

If the "Scoring Weights" section is absent or empty, the system uses the default 1.0 weights defined in modes/oferta.md. Every dimension receives equal weight, and the evaluation proceeds using the standard 10-dimension matrix without modification.

Do I need to restart Career-Ops after editing my profile files?

No. The pipeline reads modes/_profile.md at runtime during each evaluation cycle. Changes to your archetype table or scoring weights take effect immediately on the next job offer evaluation without requiring a system restart or service reload.

Can I create different weight profiles for different archetypes?

The current architecture supports a single weight profile per user in modes/_profile.md. While you can define multiple archetypes in the table, the scoring weights apply universally across all evaluations. To switch weight profiles between different career tracks, you would need to modify the weights section in modes/_profile.md before running evaluations for that specific archetype.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}