Toygar Tanyel

Toygar Tanyel

Founding AI Engineer, Promake
Visiting Researcher, Centre for Industrial Software,
University of Southern Denmark

contact.toygar at gmail dot com

I am a visiting researcher at the Centre for Industrial Software, University of Southern Denmark, in the group of Serkan Ayvaz, where I work on interpretable machine learning and language models. I studied computer science as an undergraduate at Yıldız Technical University, where one of my graduation projects was on Alzheimer's diagnosis, then did a master's in a biomedical engineering programme at Istanbul Technical University. The thesis, on counterfactual explanation for ECG, was supervised by Ilkay Oksuz and examined in computer engineering.

My curiosity started with minds rather than machines, with perception and learning and what cognitive capability is actually made of. I read the philosophy of language and human perception for their own sake and could never stop there, because the artificial version was already in front of me. How much carries over is still what I want to know, and what breaks where it does not. Under all of it is just wanting to understand things, and wanting to know what a mind is doing when it understands, which is why explanation is the problem I keep returning to.

All of it has also had to work in production. At present I design and build Promake's multi-agent stack. Earlier I built the AI stack in hBreastMMG, Hevi AI's breast-cancer detection and diagnosis product for hospitals and part of the first radiology AI suite from Turkiye that was CE marked under the MDR.

See also my full CV, the complete list of publications, my Google Scholar record, the research in more detail, and some notes on the non-technical half of this.

Email CV Scholar GitHub LinkedIn

News

Sep 2026 Promake's advertising product shipped, the second there that I have architected and built.
Jul 2026 Three papers at ICML 2026 workshops: an oral at the Philosophy of Machine Learning Workshop and two at the Workshop on Mechanistic Interpretability.
Jul 2026 Landmark regression with attention U-Net for assessing breast positioning quality in MLO mammography view published in Diagnostics.
Feb 2026 Promake's first product launched, a website builder that gets a small business onto the web.
Oct 2025 Joined Promake as founding AI engineer, with technical ownership of its agent systems.
Apr 2025 Interpretable ECG analysis for myocardial infarction detection through counterfactuals published in Biomedical Signal Processing and Control.
Mar 2025 Hevi AI's radiology suite, including hBreastMMG, the mammography product I worked on, was CE marked under the MDR, the first from Turkiye.

Publications

167 citations, h-index 7, i10-index 6 on Google Scholar, checked 15 September 2026.
† first author, * equal contribution, Q1 first-quartile journal.

  1. † Interpretable ECG analysis for myocardial infarction detection through counterfactuals

    T. Tanyel, S. Atmaca, K. Gokce, M. Y. Balik, A. Guler, E. Aslanger, I. Oksuz. Biomedical Signal Processing and Control, 2025. Q1

    doi / arXiv / code

  2. Self-reports do not identify self-models: an identifiability test for counterfactual reports

    P. M. Konrad, T. Tanyel, S. Ayvaz. Workshop on Philosophy of Machine Learning, ICML 2026. oral

    OpenReview

  3. * Landmark regression with attention U-Net for assessing breast positioning quality in MLO mammography view

    N. Denizoglu*, T. Tanyel*, E. C. Erkan, M. E. Seker, E. Meltem, et al. Diagnostics, 2026. Q1

    doi / code

  4. * Annotation-efficient, patch-based, explainable deep learning using a curriculum method for breast cancer detection in screening mammography

    O. Camurdan*, T. Tanyel*, E. Aktufan Cerekci, D. Alis, E. Meltem, N. Denizoglu, et al. Insights into Imaging, 2025. Q1

    doi

  5. † Deciphering machine-learning decisions to distinguish between posterior fossa tumor types using MRI features

    T. Tanyel, C. Nadarajan, N. M. Duc, B. Keserci. Cancers, 2023. Q1

    doi / code

  6. † Evaluation of explainability methods in medical image analysis

    T. Tanyel, I. Oksuz. In Explainable Machine Intelligence in Healthcare, 2025. book chapter

    doi

Selected on venue, contribution and field; the complete list is in this PDF.

Awards

2025 Visit grant, Danish Data Science Academy, for a research stay at the University of Southern Denmark
2025 Runner-up, ECG Challenge, British Heart Foundation Data Science Centre
2024 Third place, Artificial Intelligence in Healthcare competition, among roughly 5,000 teams
2023 Research scholarship, TUSEB, Health Institutes of Turkiye, for the ECG work behind my M.Sc. thesis
2021 Research scholarship, TUBITAK, the national research council of Turkiye, on optimising the order of training examples
2021 Third place, autonomous rover competition, Teknofest

Mentoring and teaching

since 2024 Research mentoring, University of Southern Denmark. P. Mark Konrad, an undergraduate when we started and now beginning an MSc at Tübingen; his work on mechanistic interpretability and verification became three ICML 2026 workshop papers
since 2022 Student mentoring, Yıldız Technical University. Salih Emre Karakaş, who interned at Promake and now works there part time in his final year, three other undergraduates in computer science and biomedical engineering, and project groups building AI for healthcare
2022 to 2023 Peer advisor, Yıldız Technical University. An appointed position in the computer engineering department, as the person students were directed to for academic guidance
2022 Guest lecture, Deep Learning for NLP, in Introduction to Natural Language Processing (BLM4580), Yıldız Technical University