My name is Anna Susmelj (prev. Klimovskaia). I am a machine learning researcher currently working on the design and applications of machine learning methods to computer vision and medical imaging problems. My main research interests are around various aspects of unsupervised and self-supervised learning. In particular, I focus on distribution shift robustness aspect of machine learning, such as unsupervised domain adaptation and modeling uncertainty under distribution shift. In my recent applications, I work with 3D reconstruction from sparse views with a particular focus on generative modeling and uncertainty.

I obtained my B.S. and M.S. with Honours in computational mathematics from Lomonosov Moscow State University with a minor in mathematical statistics.

I got my Ph.D. from ETH Zurich in 2018 working on the topic of causal structure learning from mass cytometry time series snapshots under supervision of Prof. Manfred Claassen.

In 2018 I did an internship at Facebook AI Research (NYC) under the supervision of Léon Bottou and Maximilian Nickel on the topic of learning hierarchical representations using hyperbolic geometry. After I joined Facebook AI Research (Paris) as a postdoctoral researcher under the supervision of David Lopez-Paz. As a postdoc, I worked on the topic of out-of-distribution generalization for the prediction of unseen drug combinations. 

Due to COVID-19 and personal reasons, I moved back to Switzerland in 2020, where I started as a Senior Data Scientist at the Swiss Data Science Center. There I worked on several projects. My main project was in the medical imaging domain: adversarial domain adaptation for optoacoustic reconstruction. The paper got an oral presentation at MIDL 2022.

In 2022 I joined an ETHZ spin-off Biognosys as Head of AI, where I was leading a small team of engineers on the applications of machine learning towards the improvement of identification and quantifications in mass spectrometry data. My role responsibilities combined defining ML strategy and projects, code implementation, and supervision of engineers and interns.

In April 2023 I started my joint postdoctoral position at ETH AI Center and Computer Vision Lab under supervision of Prof. Ender Konukoglu.

Other research activities

In 2018 I was one of the organizers of NeurIPS workshop on Causal Learning.

Since 2019 I serve as a reviewer for NeurIPS, ICML, ICLR. Since 2022 I am invited as a reviewer for TMLR and Nature Machine Intelligence.

Since 2020 I am a member of ELLIS society.