yc_analysis

YC Agent Analysis Workflow

Production-ready data analysis demonstrating PicoAgents patterns.

Key Insights

• 234/5,000+ YC companies (4.7%) now build AI agents (2024 data) • Growth: From 5 companies (2020) to 234 companies (2024) - 47x increase • Top domains: Productivity (89), Health (34), Finance (28) • Cost efficiency: 90% reduction via keyword pre-filtering

Engineering Patterns

Two-stage filtering: Keywords → AI classification saves $4+ per run Structured output: Zero hallucination with Pydantic schemas Disk checkpoints: Resume processing after interruptions Independent testing: Each step unit-testable

Quick Start

# Set credentials
export AZURE_OPENAI_ENDPOINT="https://your-resource.openai.azure.com/"
export AZURE_OPENAI_API_KEY="your-key"

# Run analysis
python workflow.py

# Run tests
python test_workflow.py

Files

• models.py - Pydantic schemas for type safety
• steps.py - Individual workflow functions (testable)
• workflow.py - Main orchestration
• test_workflow.py - Unit tests for each component
• data/ - Cache directory (gitignored)

Generated report: ./yc_analysis/data/analysis.md