Agentic Resume Builder

A Python library that helps users create professional, impact-oriented resumes through an AI-guided interactive process using Strands Agents SDK and Ollama.

Features

• AI-Guided Interview: Conversational agent that asks targeted questions to extract your professional experiences
• Impact-Oriented Content: Automatically formats content with quantifiable metrics and action verbs
• Markdown-Based: Uses Markdown as the working format for easy editing and version control
• PDF Import: Convert existing PDF resumes to Markdown for editing
• Professional PDF Output: Generate polished PDF resumes with customizable styling
• Local Processing: All data stays on your machine using Ollama for privacy
• Session Management: Save and resume work sessions at your convenience

Table of Contents

• Installation
• Quick Start
• CLI Usage
• Python API
• Configuration
• Examples
• Documentation
• Development
• Troubleshooting
• License

Installation

Prerequisites

• Python 3.9 or higher
• Ollama installed and running locally

Install Ollama

1. Visit https://ollama.ai/download
2. Follow installation instructions for your platform
3. Start Ollama: ollama serve
4. Pull a model: ollama pull llama3.2:3b

Install from source

# Clone the repository
git clone <repository-url>
cd agentic-resume-builder

# Install in development mode
pip install -e .

# Or install with development dependencies
pip install -e ".[dev]"

Install from PyPI (when published)

pip install agentic-resume-builder

Quick Start

1. Create a New Resume

# Create from default professional template
agentic-resume init my_resume.md

# Or choose a specific template
agentic-resume init my_resume.md --template academic

2. Start Interactive Session

# Let the AI guide you through building your resume (uses llama3.2:3b by default)
agentic-resume interactive my_resume.md

# Or use a different Ollama model
agentic-resume --model mistral interactive my_resume.md
agentic-resume --model llama3.2 interactive my_resume.md

The AI will ask you questions about your experiences and automatically update your resume with impact-oriented content.

3. Generate PDF

# Create a professional PDF
agentic-resume generate my_resume.md my_resume.pdf

Alternative: Import from Existing PDF

# Convert your existing PDF resume to Markdown
agentic-resume import old_resume.pdf my_resume.md

# Then continue with interactive session
agentic-resume interactive my_resume.md

CLI Usage

Global Options

These options can be used with any command:

--config PATH    # Path to custom configuration file (YAML)
--model MODEL    # Ollama model to use (default: llama3.2:3b)

Examples:

# Use a different model for all commands
agentic-resume --model mistral interactive my_resume.md
agentic-resume --model llama3.2 interactive my_resume.md
agentic-resume --model codellama interactive my_resume.md

# Use custom configuration file
agentic-resume --config my_config.yaml interactive my_resume.md

# Combine both options
agentic-resume --config my_config.yaml --model mistral interactive my_resume.md

Popular Ollama Models:

• llama3.2:3b - Default, fast and efficient (3B parameters)
• llama3.2 - Larger version with better quality (7B+ parameters)
• mistral - Alternative high-quality model
• codellama - Optimized for technical resumes
• phi - Microsoft's efficient model

To install a model: ollama pull <model_name>

Available Commands

init - Create New Resume

Create a new resume from a template.

agentic-resume init <output_path> [--template TEMPLATE]

# Examples:
agentic-resume init my_resume.md
agentic-resume init my_resume.md --template academic
agentic-resume init my_resume.md --template technical

Options:

• --template: Template type (professional, academic, technical). Default: professional

load - Load Existing Resume

Load and validate an existing Markdown resume.

agentic-resume load <markdown_path>

# Example:
agentic-resume load my_resume.md

import - Import from PDF

Import a resume from PDF format and convert to Markdown.

agentic-resume import <pdf_path> <output_path>

# Example:
agentic-resume import old_resume.pdf my_resume.md

interactive - Start AI Session

Start an AI-assisted interactive session to build your resume.

agentic-resume interactive <markdown_path> [--config CONFIG]

# Examples:
agentic-resume interactive my_resume.md
agentic-resume interactive my_resume.md --config custom_config.yaml

Options:

• --config: Path to custom configuration file

During the session:

• Answer the AI's questions about your experiences
• The AI will automatically update your resume
• Type quit or exit to save and exit
• Type pause to save and pause the session

generate - Generate PDF

Generate a professional PDF from your Markdown resume.

agentic-resume generate <markdown_path> <output_path> [--style STYLE]

# Examples:
agentic-resume generate my_resume.md my_resume.pdf
agentic-resume generate my_resume.md my_resume.pdf --style custom_style.yaml

Options:

• --style: Path to custom PDF style configuration

resume - Resume Session

Continue a previously saved interactive session.

agentic-resume resume <session_id>

# Example:
agentic-resume resume 550e8400-e29b-41d4-a716-446655440000

Note: Session IDs are displayed when you pause or exit an interactive session.

Getting Help

For detailed help on any command:

agentic-resume --help
agentic-resume <command> --help

Python API

The library provides a high-level Python API for programmatic use.

Basic Usage

from agentic_resume_builder import api

# Create a new resume
resume = api.create_resume("my_resume.md", template="professional")

# Load an existing resume
resume = api.load_resume("my_resume.md")

# Import from PDF
resume = api.import_resume_from_pdf("old_resume.pdf", "my_resume.md")

# Generate PDF
api.generate_pdf(resume, "my_resume.pdf")

Interactive Session

from agentic_resume_builder import api

# Load resume
resume = api.load_resume("my_resume.md")

# Start interactive session
session = api.start_interactive_session(resume)

# In a real application, you would present questions to the user
# and process their responses in a loop

Batch Updates

from agentic_resume_builder import api

resume = api.load_resume("my_resume.md")

# Update multiple sections at once
updates = [
    {"section_name": "Summary", "content": "Experienced software engineer..."},
    {"section_name": "Skills", "content": "Python, JavaScript, Go, AWS"}
]
resume = api.batch_update_resume(resume, updates)

# Save changes
api.save_resume(resume)

Complete Workflow

from agentic_resume_builder import api

def my_callback(question):
    """Handle questions from the AI agent."""
    print(question.text)
    return input("Your answer: ")

# Run complete workflow from start to finish
resume = api.complete_workflow(
    input_path="my_resume.md",
    output_pdf_path="my_resume.pdf",
    interactive_callback=my_callback
)

Validation and Statistics

from agentic_resume_builder import api

resume = api.load_resume("my_resume.md")

# Validate resume
validation = api.validate_resume(resume)
if not validation["is_valid"]:
    print("Errors:", validation["errors"])
    print("Warnings:", validation["warnings"])

# Get statistics
stats = api.get_resume_statistics(resume)
print(f"Completeness: {stats['completeness_score']:.1%}")
print(f"Experiences: {stats['num_experiences']}")
print(f"Metrics found: {stats['num_metrics']}")

For more examples, see the examples/ directory and API Examples README.

Configuration

You can customize the behavior using a YAML configuration file.

Configuration File Format

Create a config.yaml file:

# Ollama Configuration
ollama:
  base_url: "http://localhost:11434"  # Ollama server URL (must be local)
  model: "llama3.2:3b"                      # Model to use
  temperature: 0.7                     # Generation temperature (0.0-2.0)
  timeout: 30                          # Request timeout in seconds

# PDF Style Configuration
pdf_style:
  font_family: "Helvetica"             # Font family
  font_size: 11                        # Base font size
  heading_font_size: 14                # Heading font size
  line_spacing: 1.2                    # Line spacing multiplier
  
  # Margins (in cm)
  margin_top: 2.0
  margin_bottom: 2.0
  margin_left: 2.0
  margin_right: 2.0
  
  # Colors (hex format)
  color_primary: "#000000"             # Primary text color
  color_secondary: "#666666"           # Secondary text color
  color_accent: "#0066cc"              # Accent color for highlights

# Template and Session Directories
templates:
  directory: "./templates"
  default: "professional"

sessions:
  directory: "./sessions"
  auto_save: true
  save_interval: 60                    # Auto-save interval in seconds

Using Configuration

With CLI:

agentic-resume --config config.yaml interactive my_resume.md

With Python API:

from agentic_resume_builder import ResumeConfig

# Load from YAML
config = ResumeConfig.from_yaml("config.yaml")

# Or create programmatically
from agentic_resume_builder import OllamaConfig, PDFStyle

config = ResumeConfig(
    ollama=OllamaConfig(
        model="llama3.2:3b",
        temperature=0.8
    ),
    pdf_style=PDFStyle(
        font_family="Arial",
        font_size=11
    )
)

Configuration Options

Ollama Options

Option	Type	Default	Description
base_url	string	http://localhost:11434	Ollama server URL (must be local)
model	string	llama3.2:3b	Model name to use
temperature	float	0.7	Generation temperature (0.0-2.0)
timeout	int	30	Request timeout in seconds

Recommended Models:

• llama3.2:3b: General purpose, good balance
• llama3: Improved reasoning
• mistral: Fast and efficient
• codellama: For technical resumes
• phi: Lightweight and fast

PDF Style Options

Option	Type	Default	Description
font_family	string	Helvetica	Font family name
font_size	int	11	Base font size in points
heading_font_size	int	14	Heading font size in points
line_spacing	float	1.2	Line spacing multiplier
margin_top	float	2.0	Top margin in cm
margin_bottom	float	2.0	Bottom margin in cm
margin_left	float	2.0	Left margin in cm
margin_right	float	2.0	Right margin in cm
color_primary	string	#000000	Primary text color (hex)
color_secondary	string	#666666	Secondary text color (hex)
color_accent	string	#0066cc	Accent color (hex)

For more details, see docs/configuration.md.

Examples

The examples/ directory contains comprehensive examples:

• api_basic_usage.py - Basic API operations (create, load, import, generate)
• api_interactive_workflow.py - Interactive sessions with the AI agent
• api_advanced_patterns.py - Advanced patterns and best practices
• resume_manager_example.py - Using ResumeManager directly
• pdf_generator_example.py - PDF generation with custom styling
• ollama_client_example.py - Ollama integration examples

See examples/API_EXAMPLES_README.md for detailed documentation.

Documentation

• User Guide - Complete guide for end users
• API Reference - Detailed API documentation
• Configuration Guide - Configuration options and examples
• Ollama Integration - Ollama setup and usage
• Agent Tools Usage - Using agent tools
• Troubleshooting - Common issues and solutions
• Development Guide - For contributors

Development

Setup

# Clone repository
git clone <repository-url>
cd agentic-resume-builder

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install with development dependencies
pip install -e ".[dev]"

Running Tests

# Run all tests
pytest

# Run specific test types
pytest tests/unit/           # Unit tests only
pytest tests/property/       # Property-based tests only
pytest tests/integration/    # Integration tests only

# Run with coverage
pytest --cov=agentic_resume_builder --cov-report=html

# Run with verbose output
pytest -v

Code Quality

# Format code
black src/ tests/

# Lint code
ruff check src/ tests/

# Type checking (if mypy is installed)
mypy src/

Project Structure

agentic-resume-builder/
├── src/agentic_resume_builder/    # Main package
│   ├── __init__.py                # Package initialization & API
│   ├── models.py                  # Data models (Pydantic)
│   ├── exceptions.py              # Custom exceptions
│   ├── config.py                  # Configuration models
│   ├── manager.py                 # Resume manager
│   ├── agent.py                   # AI agent (Strands)
│   ├── agent_tools.py             # Agent tools
│   ├── markdown_processor.py      # Markdown parsing/rendering
│   ├── document_converter.py      # PDF to Markdown conversion
│   ├── pdf_generator.py           # PDF generation
│   ├── session_manager.py         # Session persistence
│   ├── ollama_client.py           # Ollama integration
│   └── cli.py                     # Command-line interface
├── tests/
│   ├── unit/                      # Unit tests
│   ├── property/                  # Property-based tests
│   ├── integration/               # Integration tests
│   └── conftest.py                # Shared fixtures
├── examples/                      # Usage examples
├── docs/                          # Documentation
├── pyproject.toml                 # Project configuration
└── README.md                      # This file

For more details, see DEVELOPMENT.md.

Troubleshooting

Ollama Connection Issues

Problem: "Could not connect to Ollama"

Solutions:

1. Ensure Ollama is running: ollama serve
2. Check if Ollama is accessible: curl http://localhost:11434/api/tags
3. Verify the port in your configuration matches Ollama's port

Model Not Found

Problem: "Model not found: llama3.2:3b"

Solutions:

1. Pull the model: ollama pull llama3.2:3b
2. List available models: ollama list
3. Use an available model in your configuration

PDF Generation Fails

Problem: "PDF generation failed"

Solutions:

1. Check that your Markdown is valid
2. Ensure weasyprint dependencies are installed (see weasyprint docs)
3. Try with default styling first
4. Check file permissions for output directory

Import from PDF Issues

Problem: "PDF conversion produced poor results"

Solutions:

1. Ensure the PDF has selectable text (not scanned images)
2. Try manually cleaning up the converted Markdown
3. Consider starting from a template if conversion quality is poor

Session Not Found

Problem: "Session not found"

Solutions:

1. Check that the session ID is correct
2. Verify the sessions directory exists and has proper permissions
3. Check if the session file was accidentally deleted

For more troubleshooting help, see docs/troubleshooting.md.

Requirements

• Python 3.9 or higher
• Ollama installed and running locally
• Sufficient disk space for models (typically 4-8 GB per model)
• Internet connection for initial model download

Privacy and Security

• All processing is local - No data is sent to external servers
• No telemetry - We don't track usage or collect data
• Session files are saved with restrictive permissions (600)
• Your resume data never leaves your machine

License

MIT License - see LICENSE file for details

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

How to Contribute

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes
4. Run tests (pytest)
5. Commit your changes (git commit -m 'Add amazing feature')
6. Push to the branch (git push origin feature/amazing-feature)
7. Open a Pull Request

Support

• Issues: Report bugs or request features via GitHub Issues
• Documentation: See the docs/ directory
• Examples: Check the examples/ directory

Acknowledgments

• Built with Strands Agents SDK
• Powered by Ollama
• PDF conversion using MarkItDown
• PDF generation with WeasyPrint