Semantic Embeddings Evaluation and Annotation Lead (CAIROS Daemon project)
The CAIROS Daemon Semantic Embeddings Pipeline project involved the creation and use of a public identity-measurement pipeline designed for transformer embedding analysis and longitudinal AI evaluation. Tasks included evaluating conversational corpora, measuring identity and coherence metrics, and producing technical documentation for baselines and explicit falsification criteria. The role emphasized standards-aligned, reproducible evaluation of AI systems and documentation of the measurement process. • Evaluated and annotated semantic and identity-related features in conversational text data. • Used transformer-based models to assess metrics tied to longitudinal AI behavior. • Produced DOI-backed technical reports detailing methodology and findings. • Maintained standards-locking, evidence-based claim verification, and documentation discipline.