EU AI Act Transparency Notice

RepodIn uses artificial intelligence (AI) to analyze code repositories, assess developer skills, and generate insights. This transparency notice explains how we use AI and your rights under the EU AI Act. What AI Systems We Use: - Code analysis AI models (Claude, GPT-4, Gemini, DeepSeek, Mistral) - Natural language processing for report generation - Pattern recognition for code quality assessment - Skills assessment algorithms for developer portfolios and skills reports Developer B2C use (skills reports, /in/[username] portfolios): - AI generates informational skills reports and portfolio summaries from code you choose to analyze - Results are shown to you first; sharing a public portfolio link is your choice - We do not make automated hiring or employment decisions on your behalf Risk Classification: RepodIn does not apply a blanket EU AI Act risk classification. Developer self-service, education and employment-related use cases are assessed separately before production use. We do not make solely automated decisions that significantly affect legal rights or opportunities in the developer segment. Education grading workflows may involve separate risk classification — see institution compliance templates for teacher use.

AI is used in the following scenarios: Code Repository Analysis: - When you request analysis of a GitHub repository - When analyzing code quality, security, and best practices - When generating skills assessment reports Skills Assessment & Developer Portfolio: - When evaluating your technical skills based on code you connect - When generating employer-facing skills reports - When building or updating your public portfolio at /in/[username] Report Generation: - When creating analysis reports - When generating insights and recommendations - When formatting results for export You will always be informed when AI is being used through clear indicators, consent flows, and this transparency notice. Developer analyses show which model was used where applicable.

Decision-Making Process: 1. Input Processing: Your code repository is analyzed using AI models 2. Pattern Recognition: AI identifies code patterns, quality metrics, and technical skills 3. Scoring: AI generates scores for correctness, completeness, style, documentation, maintainability, and security 4. Insight Generation: AI creates personalized insights and recommendations 5. Report Compilation: Results are compiled into a comprehensive report AI Models Used: - Claude (Anthropic): High-quality code analysis and reasoning - GPT-4 (OpenAI): Code understanding and pattern recognition - Gemini (Google): Fast analysis and cost-effective processing - DeepSeek: Code-specialized analysis - Mistral: EU-compliant AI processing Model Selection: We select AI models based on: - Analysis complexity and requirements - Cost optimization - Performance and speed - User preferences - Availability and reliability

Data Sources: Public GitHub Data: - Repository code and structure - Commit history and patterns - File organization and architecture - Technology stack information LinkedIn Profile Data (with consent): - Professional experience - Skills and endorsements - Education background - Work history CV/Resume Data (with consent): - Skills and qualifications - Work experience - Education details Analysis Metadata: - Analysis type and purpose - User preferences and settings - Historical analysis data (for improvement) Data Processing: - All data is processed securely in the EU - Data is anonymized where possible - Personal data is only used with your explicit consent - Data retention follows GDPR requirements

Important Limitations: Not Professional Advice: - AI analysis is for informational purposes only - Not a substitute for professional code review - Not legal, financial, or career advice - Results should be interpreted with context Accuracy Considerations: - AI models may have biases or limitations - Analysis quality depends on code complexity - Results may vary between different AI models - Some edge cases may not be detected Accuracy Metrics: - Code quality scores: ±5-10% variance possible - Skills assessment: Based on code patterns, not comprehensive evaluation - Recommendations: Suggestions based on best practices, not guarantees Continuous Improvement: - We regularly update AI models - We monitor and improve accuracy - We incorporate user feedback - We track model performance metrics

Right to Information: - You have the right to know when AI is being used - You can request information about AI decision-making - You can access transparency notices (this page) Right to Explanation: - You can request an explanation of how AI made a decision - We will provide clear, understandable explanations - You can access AI decision logs for your analyses Right to Human Review: - You can request human review of AI-generated results - We provide manual review options for critical analyses - You can challenge AI decisions and request reconsideration Right to Opt-Out: - You can opt-out of AI-powered analysis (with limitations) - Some features require AI and cannot be disabled - You can request deletion of AI-generated data Right to Data Access: - You can access all your data, including AI analysis results - You can export your data in standard formats - You can request data deletion at any time

What We Do: Clear Indicators: - AI-powered badges and labels on analysis pages - Transparency notices before analysis starts - Model information displayed in results - Clear explanation of AI usage Documentation: - Complete AI system documentation - Model specifications and capabilities - Decision-making process documentation - Accuracy and limitation disclosures Audit Trails: - All AI usage is logged - Decision explanations are stored - User interactions are tracked - Compliance logs are maintained User Education: - AI system user guide available - Help documentation and FAQs - Support for understanding AI results - Regular updates on AI improvements

EU AI Act Compliance: Risk Management: - Regular risk assessments - Compliance monitoring - Safety measures and safeguards - Error handling and fallbacks Quality Assurance: - Model performance monitoring - Accuracy tracking and improvement - User feedback integration - Continuous testing and validation Data Protection: - GDPR-compliant data processing - Secure data storage and transmission - Access controls and authentication - Regular security audits Human Oversight: - Human review available on request - Manual override options - Quality control processes - Escalation procedures for issues

Questions About AI Usage? General Inquiries: Email: support@repodin.com Response time: 48 hours AI Decision Explanations: Use the "Explain AI Decision" button in your analysis results Or contact: ai-support@repodin.com Human Review Requests: Use the "Request Human Review" feature in your dashboard Or contact: support@repodin.com Compliance Questions: Email: compliance@repodin.com For EU AI Act compliance inquiries Data Access Requests: Use the data export feature in Settings Or contact: privacy@repodin.com