me@dsmn.dev — bash
me@dsmn.dev:~$ whoami
Mihai Dusmanu
Mihai Dusmanu
Member of Technical Staff
Anthropic — Multimodal Pretraining
me@dsmn.dev:~$ cat about.txt

At Anthropic, making Claude better at understanding multimodal data.

Previously at Microsoft, worked on a large-scale cloud system for spatial intelligence, and later prototyped AI-powered multimodal retrieval and Q&A.

PhD from ETH Zurich on machine learning for 3D reconstruction and visual localization.

me@dsmn.dev:~$ ls ~/papers --highlight
S. Deb Sarkar, R. Pautrat, O. Miksik, M. Pollefeys, I. Armeni, M. Rad*, M. Dusmanu*
arXiv 2025
H. Blum, A. Mercurio, J. O'Reilly, T. Engelbracht, M. Dusmanu, M. Pollefeys, Z. Bauer
CVPR 2025
L. Bösiger, M. Dusmanu, M. Pollefeys, Z. Bauer
ECCV 2024
P.-E. Sarlin*, M. Dusmanu*, J. L. Schönberger, P. Speciale, L. Gruber, V. Larsson, O. Miksik, M. Pollefeys
ECCV 2022
Privacy-Preserving Image Features via Adversarial Affine Subspace Embeddings
M. Dusmanu, J. L. Schönberger, S. N. Sinha, M. Pollefeys
CVPR 2021 Best Paper Candidate · arxiv · github
D2-Net: A Trainable CNN for Joint Detection and Description of Local Features
M. Dusmanu, I. Rocco, T. Pajdla, M. Pollefeys, J. Sivic, A. Torii, T. Sattler
CVPR 2019 · arxiv · github
me@dsmn.dev:~$ cat links.txt
me@dsmn.dev:~$
built with claude code
click anywhere to skip