Dr. Hüseyin Oylupinar is a historian and AI practitioner specializing in the integration of artificial intelligence into social science, cultural memory, and historical research. His work focuses on building ethical, multilingual AI systems that recover, analyze, and preserve structurally excluded narratives—ranging from Holocaust and Roma testimonies to Ukrainian wartime oral histories and under-resourced languages such as Kalderash Romani.
Dr. Oylupinar has developed and deployed custom pipelines across the full AI stack, including OCR (Kraken), ASR (Whisper), NMT (Hugging Face Transformers), image super-resolution (Real-ESRGAN), and PyTorch-based fine-tuning for historical and literary domains. His methodological innovations span diachronic printed text, early 20th-century handwriting (including German Kurrent), and both typewritten and poetic document analysis.
His current frontier is AI-assisted literary translation, where he is building a custom neural model for Ukrainian–English translation of metaphor-rich, culturally dense literary texts. In this project, Dr. Oylupinar has advanced a number of critical skills: he fine-tunes literary-AI models using curated, high-fidelity translation pairs that preserve metaphor, tone, and narrative structure; he designs contrastive learning inputs by creating adversarial “near-miss” pairs that train the model to distinguish fine-grained semantic nuances; and he applies rigorous model failure diagnostics by identifying and correcting issues such as metaphor flattening, idiomatic drift, and emotional tone loss through targeted data design. His work also includes constraint-based output control to maintain poetic structure, verse alignment, and theological accuracy, and domain-specific evaluation to benchmark aesthetic and interpretive quality in ways that go beyond standard BLEU or ROUGE metrics. Alongside these technical achievements, he is developing AI pedagogy for the humanities, translating his methods into a hands-on workshop at the 2025 Translating Ukraine Summer Institute (Harvard and University of Wrocław), where he trains scholars to understand, fine-tune, and apply literary AI tools.
He also serves as the lead AI architect for the Horizon Europe RECUPERATE project (in application), designing a modular, explainable AI system for testimony-based peacebuilding. His work in this domain integrates multilingual embeddings, theme classification, emotion recognition, and human-in-the-loop verification, all aligned with postcolonial and memory studies methodologies. He ensures methodological transparency and architectural modularity across a 30-member international consortium of scholars, artists, and engineers.
Dr. Oylupinar regularly presents at global academic venues, including the Canadian Association of Slavists and Harvard’s Translating Ukraine Summer Institute, and is the founding director of the Institute for Knowledge, Research, and Society—a Canadian NGO committed to bridging AI and the humanities.
His research exemplifies a rare fusion of technical mastery, cultural sensitivity, and structural clarity—defining new territory for AI as a tool of historical justice, literary preservation, and interpretive collaboration.