resume-project-writer

Files (1)

• SKILL.md

SKILL.md (first lines)

# Resume Project Writer

## Purpose

Turn raw descriptions of technical projects into concise, technically accurate, and strong resume bullet points.

The goal is to communicate what the user actually built, how they built it, and what result they achieved without exaggerating their experience.

## When to use

Use this skill when the user wants to:

* describe a project on a resume;
* improve existing project bullet points;
* adapt a project description for an internship;
* emphasize ML, backend, data, or software engineering experience;
* shorten a long project description;
* make technical achievements easier for recruiters to understand;
* select the most valuable details from a project.

## Workflow

### 1. Understand the project

Extract the important facts:

* What problem was solved?
* What did the user personally implement?
* Which technologies were used?
* Which algorithms or models were used?
* How was the solution evaluated?
* What measurable result was achieved?
* What engineering decisions were important?

Do not start rewriting until the technical contribution is understood.

### 2. Identify the strongest signals

Prioritize information that demonstrates:

* technical depth;
* practical implementation;
* measurable results;
* experimentation;
* engineering responsibility;
* understanding of validation and evaluation;
* performance improvements;
* real-world constraints.

Remove details that do not help evaluate the candidate.

### 3. Build each bullet around a contribution

Prefer the structure:

**Action + technical approach + result**

For example:

> Built a time-series forecasting pipeline using LightGBM with lag and calendar features, evaluated with walk-forward validation and achieved an RMSLE of X.

Another valid structure is:

**Problem + implementation + impact**

Use whichever reads more naturally.

### 4. Make technologies specific

Prefer:

> LightGBM, pandas, scikit-learn

instead of:

> machine learning technologies

Prefer:

> walk-forward cross-validation

instead of:

> tested the model carefully

Use concrete technical terms when they accurately describe the work.

### 5. Include metrics when available

Good metrics include:

* validation score;
* accuracy;
* F1;
* ROC-AUC;
* RMSLE;
* RMSE;
* latency;
* memory usage;
* throughput;
* dataset size;
* training speed;
* percentage improvement.

Never invent metrics.

If no quantitative result exists, describe the technical outcome instead.

### 6. Adapt to the target role

For ML positions, prioritize:

* datasets;
* feature engineering;
* validation;
* models;
* metrics;
* experiments;
* data leakage prevention;
* model comparison.

From Antichnoste/ai-skills, MIT.

More from Antichnoste/ai-skills