Dima Pozdeev

Hi :)

I am curious about applications and fundamentals of machine learning.

Currently working on recurrent memory mechanism for video data.

I studied computer science/applied math at TUM (MS) and at HSE (BS).

Feel free to reach out on X or via email if you want to chat.

           

profile photo

Research
DenseMarks: Learning Canonical Embeddings for Human Heads Images via Point Tracks
Dmitrii Pozdeev, Alexey Artemov, Ananta R. Bhattarai, Artem Sevastopolsky
ICLR, 2026
project page / arXiv / video / code

Learning per-pixel 3D embeddings from in-the-wild head videos improves dense correspondences across subjects and poses, and boosts monocular head tracking.

To Stay or Not to Stay in the Pre-train Basin: Insights on Ensembling in Transfer Learning
Ildus Sadrtdinov*, Dmitrii Pozdeev*, Dmitry Vetrov, Ekaterina Lobacheva
NeurIPS, 2023 (* equal contribution)
arXiv / X thread / code

Exploring the target task's loss landscape in the transfer learning setup lets you build better ensembles (and soups).

GG-Langevin teaser Generative Shape Reconstruction with Geometry-Guided Langevin Dynamics
Linus Härenstam-Nielsen, Dmitrii Pozdeev, Thomas Dagès, Nikita Araslanov, Daniel Cremers
WACV, 2027
code / arXiv

Traversing diffusion model trajectories while enforcing measurement consistency at each step enables complete and realistic 3D shape reconstruction from partial observations.