Welcome!

Hello! I’m Aylin, a graduate student at ETH Zurich, currently working on my thesis with Prof. Ludwig Schmidt’s group at Stanford on generating open-source datasets for computer use models.

Latest blog post

• Latest: How Confident Are You, ChatGPT?

  TL;DR: The key to OpenAI’s IMO gold likely wasn’t a universal verifier. Calibration and uncertainty estimation are central, and will shape future products.

  Read more →

Projects

• IDE Grounding Kit: I developed a tool for automatically generating agent training data for the Cursor IDE using its underlying Electron Browser.

Personal

In my free time I love to explain my research ideas to my (probably tired from coding) boyfriend who is also an AI researcher! When I’m not immersed in the world of large language models, I enjoy learning about other cultures in ridiculous detail and experimenting with their culinary techniques, particularly those involving koji and fermentation like in this Noma fermentation guide or preparing a comprehensive reading class for Dostoevsky’s Brothers Karamasov.

Publications

• EasyARC: Evaluating Vision Language Models on True Visual Reasoning. ViSCALE Workshop @ CVPR 2025. arXiv

Feel free to reach out if you share similar interests or have any questions!