Estimating Software Team Compatibility via Behavioral Telemetry and Adaptive Psychometric Profiling

This paper presents GitSyntropy, a system for estimating software team compatibility using behavioral telemetry and adaptive psychometric profiling. The framework integrates GitHub-derived activity signals (commit timing, pull request behavior) with an 8-item computerized adaptive testing (CAT) instrument to generate pairwise compatibility scores across eight behavioral dimensions.

A circular-coordinate K-Means algorithm is used to infer developer chronotypes from commit timestamps, addressing the midnight boundary problem in temporal data. These signals are combined in a weighted compatibility model (max score: 36), and a Monte Carlo simulation module estimates optimal candidate profiles for improving team compatibility.

Evaluation on 46 real GitHub developer profiles (10,886 commits) shows strong separation between chronotype-aligned and mismatched pairs (Cohen’s d = 3.71, p < 0.001). The CAT module reduces assessment length by 37.5% while maintaining high score fidelity (r = 0.965).

Results reflect internal model behavior and require validation against real-world team performance outcomes.