Verena Kaynig

Also known as Verena Kaynig-Fittkau, Verena S. Kaynig-Fittkau

Verena Kaynig is a machine learning scientist whose work moved from reconstructing neurons in electron microscopy images to document parsing at Adobe. She earned a diploma in computer science at the University of Hamburg in 2006 and a PhD at ETH Zurich in 2011, then worked at Harvard as a postdoctoral fellow in Hanspeter Pfister's GVI Lab and, from 2014 to 2017, as a lecturer who taught CS109 data science. Her first-author paper on large-scale automatic reconstruction of neuronal processes appeared in Medical Image Analysis in 2015, and she coauthored the Trainable Weka Segmentation tool in 2017. After joining Adobe in 2017, she won its Tech Excellence Award in 2019 and coauthored the DocTime parser in 2022. She is a Senior Machine Learning Scientist II at Adobe.

Now at: Adobe
Works: 29
Citations: 70,525
h-index: 14
Badges: Bridge

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

2006–2011

Hamburg diploma and ETH PhD

She receives her diploma in computer science from the University of Hamburg in 2006. She completes a PhD in computer science at ETH Zurich in 2011, working on connectomics in Joachim M. Buhmann's Pattern Analysis and Machine Learning Group. Her open-source lens distortion correction for transmission electron microscope images is part of Fiji.
2011–2014

Postdoc in Pfister's lab

She joins Harvard's School of Engineering and Applied Sciences as a postdoctoral fellow in 2011, working in Hanspeter Pfister's GVI Lab.
2013–2015

Reconstructing neurons at scale

In 2013, she posts a first-author preprint on large-scale automatic reconstruction of neuronal processes from electron microscopy images, written with Hanspeter Pfister, Jeff Lichtman and colleagues; the paper appears in Medical Image Analysis in 2015. Its pipeline trains a random forest classifier on sparse user annotations, scales to datasets in the gigabyte-to-terabyte range, and comes with Mojo, a proofreading tool.
2014–2017

Lecturer for Harvard's CS109

She is a lecturer at Harvard from 2014 to 2017. In 2014, she is a computer science instructor of the CS109 Data Science course with Rafael Irizarry; in 2016, she teaches CS109B, Advanced Topics in Data Science, with Hanspeter Pfister and Joe Blitzstein.
2017–2018

Weka segmentation and proofreading

In 2017, she coauthors Trainable Weka Segmentation in Bioinformatics, a machine learning tool that trains a classifier on a limited number of manual annotations and segments the remaining microscopy data automatically. In 2018, she coauthors the CVPR paper on guided proofreading for connectomics, whose convolutional neural networks recommend candidate merges and splits in automatic segmentations.
2019

Adobe's Tech Excellence Award

After joining Adobe in 2017, she is one of four winners of the inaugural Tech Excellence Award in 2019, a peer-nominated award, as a Sr. Machine Learning Scientist in Digital Media.
2022

DocTime at NAACL 2022

In 2022, she is a coauthor of DocTime, a document-level temporal dependency graph parser that Adobe presents at the NAACL 2022 conference.
As of October 2026

Senior scientist at Adobe

She continues at Adobe as a Senior Machine Learning Scientist II.