PeopleofSuperintelligence

Researcher

Cengiz Oztireli

Also known as Cengiz Öztireli

Cengiz Oztireli is a computer scientist whose AI research connects visual computing with neural network attribution and brain visual decoding. He completes a double major in computer engineering and electronics engineering at Koç University and earns MS and PhD degrees in computer science from ETH Zurich; before Cambridge, he holds research posts at ETH Zurich and Disney Research. He received a UKRI Future Leaders Fellowship in 2020 for 3D digital reality research. At SIGGRAPH in August 2025, he lectured in a course on randomness in generative AI while affiliated with Cambridge and Google; his brain visual decoding paper appeared as an oral presentation at AAAI 2026. As of October 2026, he is a professor in Cambridge's Department of Computer Science and Technology and directs Core Lab.

Works
140
Citations
3,771
h-index
28
Badges

Publication numbers from OpenAlex, as of 27 October 2026.

Timeline

  1. 2017–2018

    Sensitivity-n and attribution

    He coauthors Towards better understanding of gradient-based attribution methods for Deep Neural Networks, which appears on arXiv in 2017 and at ICLR in 2018. The paper compares four gradient-based attribution methods, introduces Sensitivity-n, and tests attribution methods on image and text classification datasets.

  2. By 2020

    Education and research posts

    Before Cambridge, he works as a senior research associate at ETH Zurich and a research scientist at Disney Research; by 2020, he is a Senior Lecturer at Cambridge.

  3. October 2020

    A UKRI fellowship

    In October 2020, he receives a UKRI Future Leaders Fellowship for research on 3D digital reality and machine learning/AI.

  4. 2025–2026

    Brain visual decoding

    He and Weihao Xia develop BASIC, a framework that evaluates brain visual decoding through structural fidelity, inferential alignment and contextual coherence. The paper appears on arXiv in July 2025 and is an oral paper at AAAI 2026.

  5. August 2025

    Generative AI at SIGGRAPH

    At SIGGRAPH in August 2025, he is a lecturer for Demystifying Noise: Role of Randomness in Generative AI, a course about noise in generative modeling, with affiliations at Cambridge and Google.

  6. As of October 2026

    Professor and lab director

    He is Professor of Visual Computing and Artificial Intelligence at the University of Cambridge, Research Scientist Manager at Google, and director of Core Lab. He is principal lecturer for Cambridge’s 2026–27 Machine Visual Perception course.

Career

  1. NowUniversity of CambridgeProfessor of Visual Computing and Artificial Intelligence
  2. NowGoogleResearch Scientist Manager
  3. NowUniversity of CambridgeDirector of Core Lab
  4. ETH ZurichSenior Research Associate
  5. The Walt Disney CompanyResearch Scientist
  6. University of CambridgeSenior Lecturer

Education

  1. Koç UniversityDouble major in computer engineering and electronics engineering
  2. ETH ZurichMS in computer science
  3. ETH ZurichPhD in computer science

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Checked career sources: sites.google.com.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Cengiz Oztireli yet. Send us a free-licensed photo.

  1. Towards better understanding of gradient-based attribution methods for Deep Neural Networks arxiv.org. Checked 27 October 2026.
  2. Home sites.google.com. Checked 27 October 2026.
  3. Future Leaders Fellowship for 3D digital reality researcher cst.cam.ac.uk. Checked 27 October 2026.
  4. Multigranular Evaluation for Brain Visual Decoding arxiv.org. Checked 27 October 2026.
  5. Presentation - SIGGRAPH 2025 Conference Schedule s2025.conference-schedule.org. Checked 27 October 2026.
  6. Department of Computer Science and Technology – Course pages 2026–27: Machine Visual Perception cl.cam.ac.uk. Checked 27 October 2026.
  7. Team – Core Lab v2.core-lab.io. Checked 27 October 2026.