Home → About Us → Leadership → Dr. Hüseyin Oylupinar → AI Models and Applications Developed by Hüseyin Oylupinar
Dr. Hüseyin Oylupinar develops applied AI models and research pipelines for historical, cultural, linguistic, and memory-focused projects. His work combines custom OCR/HTR, speech recognition, neural translation, image enhancement, and human-in-the-loop review to support the recovery, transcription, translation, and interpretation of historically marginalized or difficult-to-access sources.
Selected Models and Applications
Custom AI Workflow for Source Recovery, Transcription, and Translation
Interiano 1502: Italian Source Transcription and English Translation
A custom AI-assisted transcription and translation workflow for Giorgio Interiano’s Vita et Sito de Zychi, chiamati Ciarcassi (Venice, 1502), one of the earliest printed European accounts concerning the Zychi/Circassians. The workflow uses the original Italian text and critical transcription as its source base, producing a literal English translation while keeping historical terminology and interpretation separate from the translated text.
[See it here]
Historical OCR/HTR Pipelines
Custom OCR and handwriting-recognition workflows for early printed texts, archival documents, German Kurrent handwriting, typewritten materials, and multilingual historical sources. These pipelines support transcription, correction, metadata preparation, and translation-ready source production.
[Add link if available]
Multilingual Testimony Processing
AI-assisted workflows for testimony-based collections, including transcription, translation, speaker-aware processing, thematic classification, and human verification. These applications are designed for ethically sensitive historical materials, including Holocaust, Roma, Ukrainian, and wartime oral-history contexts.
[Add link if available]
Literary AI Translation Models
Custom neural translation models for metaphor-rich, culturally dense literary texts, with emphasis on preserving tone, imagery, narrative structure, and interpretive nuance beyond standard machine-translation metrics.
[Add link if available]
Under-Resourced Language AI Workflows
Applied AI pipelines for languages and textual traditions with limited digital infrastructure, including Kalderash Romani and other historically under-resourced linguistic contexts.
[Add link if available]