David Soergel

I'm currently working on making machine learning accessible to a wide audience through clean and usable TensorFlow APIs. I'm particularly interested in easing burdens associated with data management, data preprocessing, and provenance tracking--which in practice can often require more user code and engineer time than the machine learning algorithms themselves. Previously: Ph.D. in Biophysics (Berkeley 2010); founder of OpenReview.net; scientific software quality cynic; etc.

Google Publications

Previous Publications

  •   

    Rampant software errors may undermine scientific results

    David Soergel

    F1000 Research, vol. 3 (2015), pp. 303

  •   

    Open Scholarship and Peer Review: a Time for Experimentation

    David A W Soergel, Adam Saunders, Andrew McCallum

    ICML 2013 Workshop on Peer Reviewing and Publishing Models (2013)

  •   

    Selection of primers for optimal taxonomic classification of environmental 16S rRNA gene sequences

    David A W Soergel, Neelendu Dey, Rob Knight, Steven E Brenner

    The ISME Journal, vol. 6 (2012), pp. 1440-1444