Grigorii
Khvatskii.
Also Gregory, Grigoriy, Grisha, and Гриша.
Transliteration is complicated ↗
My research focuses on using Large Language Models (LLMs) in low-resource domains.
Previously, I studied and worked at the Higher School of Economics, doing research in computational social science and teaching R and Python for data science.
Away from research, I write about Linux, experiment with code, and automate the boring things. Occasionally, cooking enters the picture.
More about my backgroundFrom my research
Why AI Detection Fails for Academic Integrity
AI detectors can flag harmless edits and miss disguised AI writing, so their scores should not be treated as proof of cheating.
Jonathan A. Karr Jr, Grigorii Khvatskii, Ting Hua & Nitesh V. Chawla
arXiv preprint · arXiv:2608.11256Notes & experiments
Getting Chromium-based browsers to a usable state on Wayland
Hardware acceleration, Mesa workarounds, and smoother scrolling on Linux.
Running RStudio Server in a rootless Podman container
A practical setup for fast R packages with RStudio Server and rootless Podman.
FizzBuzz from scratch or life without natural numbers
Building natural numbers from scratch, then taking the scenic route to FizzBuzz.