Daniel Paysan

PhD Student at ETH Zurich and the Paul Scherrer Institute | Machine Learning, Computational Biology


ETH Zurich

July 2020 — June 2024 (expected)

Doctor of Science ETH, Computational Biology and Machine Learning

Zurich, Switzerland

  • Specialization in biomedical, multi-scale and multi-modal data integration, image analysis and comuter vision
  • Thesis: Machine Learning Approaches Linking Gene Regulation and Genome Organization in Health and Disease

ETH Zurich

September 2017 — May 2020

Master of Science ETH, Statistics; GPA: 5.71/6.00

Zurich, Switzerland

  • Specialization in data science and AI through elective course work, volunteer semester projects and research assistantships
  • Thesis: Self-Supervised Representation Learning for Surgical Action Recognition (Grade: 6.00/6.00)

Berlin School of Economics and Law

October 2014 — September 2017

Bachelor of Science, Information Systems, GPA: 1.4 (ECTS grade A)

Berlin, Germany

  • Integrated degree program with IBM Germany which combined academic studies at a university of applied sciences with work experience in the industry via various internships at IBM including research and development, internal and client-facing roles
  • Thesis: Analysis of Machine Learning Algorithms for Systems` Performance Anomaly Detection (optimal grade)

Selected Experience

Paul Scherrer Institute

July 2020 — Present

Ph.D. Candidate in Computational Biology and Machine Learning

Villigen, Switzerland

  • Developed various computational methods and machine learning pipelines using imaging and multiomic data to identify novel diagnostic biomarkers; Pipelines were primarily implemented using Pyton and PyTorch
  • Co-authored 5 manuscripts and collaborated with clinical experts and researchers from various fields

ETH Zurich

May 2020 — July 2020

Research Assistant

Zurich, Switzerland

  • Developed and published a framework to semi-automatically extract interpretable features from surgical video data by combining self-supervised deep representation learning and statistical time series analyses
  • The use of the extracted features in combination with sensor data enabled up to six times more accurate unsupervised recognition of surgical activities compared to the setting where only sensor data was used

ETH Zurich

February 2019 — September 2019

Research Assistant

Zurich, Switzerland

  • Developed a hidden semi-Markov model for unsupervised surgical activity recognition from sensor data enabling the automatic and up to 50x faster (compared to the manual) annotation of complex surgical workflows
  • Built a simple implementation of the model in Python enabling end-to-end training and inference

IBM Research & Development

May 2017 — September 2017

Research Intern

Boeblingen, Germany

  • Constructed a machine learning pipeline in Python using ensemble models to identify systems performance anomalies for large-scale mainframe systems with up to 95% accurracy
  • Developed several interactive dashboards to visualize the analyses results using Apache Zeppelin and Spark and summarized the results in the Bachelor`s thesis

IBM Global Business Sevices

November 2016 — February 2017

IT Architecture Intern

Teltow, Germany

  • Built a Java software package enabling the automatic benchmarking of the performance of the OCR component of an ECM system of a leading company
  • Designed multiple SQL databases and Python scripts to automate the software deployment and generate project dashboards as part of a team of 10+ IT architects and software developers

IBM Research

May 2016 — August 2016

Research Intern

San Jose (CA), United States

  • Developed a statistical time series model in R simulating te glucose level of diabetics tereby accounting for the daily schedules of the individuals
  • Built a prototype of a data processing system using Apache Spark and Java that predicts future hypoglycemic events from streaming glucose sensor data in near-real time

IBM Global Business Services

January 2016 — February 2016

Software Develpment Intern

Hanover, Germany

  • Developed a software module in Java simulating key peripheral system components of an ECM system of a global business insurance company
  • The module enabled system-wide API testing up to 30% earlier during the software development and testing cycle for the 20+ developers of the project team

Additional Experience


Oral presentation at the International Conference of Systems Biology (October 2023, Hartford, USA) with the title Image2Reg: Linking Chromatin Images to Gene Regulation using Gene Pertubation Screens


One of the two IT supporters of the Laboratory of Nanoscale Biology at the Paul Scherrer Institute with 45+ associated staff members; Teaching Assistant for a graduate course with 55+ students

Peer Review

Peer reviewer for several scientific conferences and journals including RECOMB, Scientific Reports, BMC Bioinformatics and Signal, Image and Video Processing