Member of Technical Staff at Krea, working on generative computer vision. Previously Kandinsky Labs, VILAB, EPFL and Samsung AI Center. Based in Dubai.
- Krea 2 — open-weights text-to-image foundation model trained from scratch; ranked second worldwide on style fidelity
- XLabs FLUX LoRAs — the first style adapters for FLUX.1-dev; 500k downloads in the first month
- Kandinsky 3.0 / 3.1 / 4.0 — text-to-image and text-to-video foundation models; co-author of the EMNLP 2024 paper
- Task Discovery — NeurIPS 2022, finding the tasks neural networks generalize on
- Age of Autoresearch — expertise moves from "find the solution" to "build the system that finds it" (RU/EN)
- GPU Story — how a triangle-drawing chip became the most important processor on the planet (RU/EN)
- Realtime Video Generation Calculator — first-principles compute estimator for streaming video models
Website · Google Scholar · X · LinkedIn · Telegram (RU)