Exploring the intersection of artificial intelligence, natural language processing, and classical scholarship
A complete pipeline for processing 19th century Greek diary texts using Claude API for error correction and UGARIT for named entity recognition. Reduces character error rates from 7% to 2-3% at ~$0.50 per 750-page document.
A platform for collecting and trading stickers with other enthusiasts.
This site showcases experimental projects combining modern AI techniques with classical scholarship and digital humanities. The focus is on practical applications of large language models, NER systems, and data processing pipelines for historical text analysis.