Methodology
Principles
AURA is a public engineering signal platform. We analyze publicly available GitHub data to create a multidimensional representation of a developer's visible engineering aura.
This is not a measure of developer worth, intelligence, or capability. It is a versioned interpretation of observable public signals.
Dimensions
Shipping
Consistency and frequency of code delivery. Measures repo count, recency of activity, and account longevity.
Architecture
Complexity and structure of repositories. Considers repo size, documentation, originality, and templates.
Craft
Code quality signals. Analyzes stars, forks, language diversity, and modern practices (topics).
Collaboration
Community engagement. Measures followers, fork adoption, and issue engagement.
Signal
Social proof and visibility. Considers network size, impact metrics, and profile completeness.
Specialization
Depth of expertise. Analyzes language concentration and primary domain expertise.
Versioning
All scores are versioned using semantic versioning (e.g., v1.0.0). When the scoring model is updated, all profiles are re-analyzed and assigned the new version number.
Current version: v1.0.0
Confidence
Confidence scores (0-100%) indicate the richness of available public data. Higher confidence means more repositories were analyzed and more signals were available.
Low confidence does not mean low skill—it means limited public signal.
Transparency
We disclose our principles, dimensions, and methodology. However, exact scoring weights and implementation details remain proprietary.
This balance ensures the model cannot be easily gamed while maintaining transparency about what we measure.