Curriculum Vitae

You can download my full CV as a PDF

Education

PhD in Cognitive Science — METU, Ankara (Sep 2024 – Present)

MSc in Cognitive Science — METU, Ankara (Oct 2020 – May 2024)
Thesis: Transformer Models for Translating Natural Language Sentences into Formal Logical Expressions

BSc in Computer Engineering — Koç University, İstanbul (Sep 2014 – Jun 2020)
Track Program Certificate in Brain and Cognitive Sciences

Research Experience

• UNI-LM: A Universal Framework for Language Modeling (BİDEB-2232 / Lider Araştırmacılar Programı), Ankara, Turkey (Jun 2025 – Present)
  • Investigating the intersection of cognitive science and AI by integrating interactionist theories of reasoning into multi-agent systems.
  • Designing and implementing RL-based training pipelines and dense reward functions to enhance multi-agent reasoning, argumentation, and model generalization.
• Spatial Audio Signal Processing for 6-DoF Immersive Environments (TÜBİTAK 1001 Project), Ankara, Turkey (Mar 2022 – Jul 2023)
  • Worked as a TÜBİTAK graduate scholar under Hüseyin Hacıhabiboğlu, PhD, developing virtual reality environments and 3D object models using Reality Capture for spatial sound processing experiments and algorithm testing in Unity.
• Koç University Research in Applied Memory (KURAM), İstanbul, Turkey (Jan 2018 – Jun 2020)
  • Worked as an undergraduate research assistant under Sami Gülgöz, PhD, assisting in designing and running performance tasks, collecting and analyzing behavioral data, and receiving training in experimental design.

Publications

• Deveci, İ. E., Tan Çalık, F., Sağlam, B. D., & Ataman, D. (2026). ArgGYM: A procedural, engine-verified benchmark for structured defeasible reasoning [Preprint]. arXiv. arXiv · GitHub

• Deveci, İ. E. (2026). A Diagnostic Framework for Compositional Generalization via NL-to-FOL Translation. Journal of Logic, Language and Information. doi:10.1007/s10849-026-09480-0

• Olgun, O., Deveci, İ. E., & Hacıhabiboğlu, H. (2026). Parametric sound field interpolation for scene-based navigable immersive audio. IEEE Transactions on Audio, Speech and Language Processing, 34, 1608–1620. doi:10.1109/TASLPRO.2026.3671646

• Deveci, İ. E. (2025). Yapay zekada sembolik ve sembolik olmayan ayrımı üzerine [On the distinction between symbolic and subsymbolic in artificial intelligence]. In Z. Göz (Ed.), Yapay zekâ felsefesi (pp. 173–236). Doruk Yayınları. Doruk Yayınları

• Deveci, İ. E., & Ataman, D. (2025). The Ouroboros of benchmarking: Reasoning evaluation in an era of saturation. NeurIPS 2025 Workshop on Evaluating the Evolving LLM Lifecycle: Benchmarks, Emergent Abilities, and Scaling. NeurIPS · OpenReview · arXiv

• Chang, T. A., Arnett, C., Eldesokey, A., … Deveci, İ. E., … Wang, J. (2025). Global PIQA: Evaluating physical commonsense reasoning across 100+ languages and cultures [Preprint]. arXiv. arXiv

• Deveci, İ. E. (2025). Transformer models for translating natural language sentences into formal logical expressions. In V. Gradimondo & E. E. Rosina (Eds.), Proceedings of the ESSLLI 2025 Student Session, 36th European Summer School in Logic, Language, and Information (pp. 162–171). PhilArchive

• Deveci, İ. E., Küskü, M., & Özkurt, T. E. (2021). Makinenin beni ve ötekisi [Machine’s self and other]. Pasajlar Sosyal Bilimler Dergisi, 9 (Yapay Zekâ).

Conference Presentations

• Deveci, İ. E. (2025, July 1-3). Evaluating Compositionality in Large Language Models through Natural Language to First-Order Logic Translation [Poster presentation]. Joint IACAP/AISB Conference on Philosophy of Computing and AI, University of Twente, Enschede, Netherlands.

• Deveci, İ. E. (2024, December 6). Why Machine Learning Needs Task-specific Metrics: A Case Study on Semantic Parsing Models [Poster presentation]. METU Open Research Day, Ankara, Turkey.

• Deveci, İ. E. (2024, June 1). Transformer Models for Translating Natural Language Sentences into Formal Logical Expressions [Poster presentation]. ISBCS 2024, Ankara, Turkey.

• Deveci, İ. E., Zambak, A. F., & Temürcü, C. (2023, December 13). Bridging Words and Symbols: Using Transformers for Translating Sentences into Logical Expressions [Poster presentation]. METU Open Research Day, Ankara, Turkey.

• Deveci, İ. E., & Başoğlu, C. (2023, May 7). The Main Problem with Neuroimaging Research in Explaining Cognitive Faculties [Poster presentation]. ISBCS 2023, Istanbul, Turkey.

• Şahin, Y., & Deveci, İ. E. (2023, May 7). Against Representations That Do Not Represent: Reply to Thomson and Piccinini [Poster presentation]. ISBCS 2023, Istanbul, Turkey.