cv

Basics

Name Katarina Elez
Label Researcher
Email katarina.elez1@gmail.com
Url https://katarinaelez.github.io/
Summary Machine learning researcher in drug discovery; 6+ years of experience in molecular machine learning and large-scale molecular simulations.

Education

  • 2019.11 - Present

    Berlin, Germany

    PhD
    Free University of Berlin
    Bioinformatics
  • 2017.10 - 2019.07

    Bologna, Italy

    MSc
    University of Bologna
    Bioinformatics
    • Grade: 110/110 cum laude
    • Thesis: Development of a deep learning method for model quality assessment
  • 2013.10 - 2016.12

    Bari, Italy

    BSc
    University of Bari
    Computer Science
    • Grade: 110/110 cum laude
    • Thesis: A novel approach for liver and hepatocellular carcinoma segmentation from triphasic CT images

Work

  • 2025.02 - 2025.03

    Eindhoven, Netherlands

    Guest Researcher
    Eindhoven University of Technology, group of Prof. Grisoni
  • 2024.11 - 2024.12

    Eindhoven, Netherlands

    Guest Researcher
    Eindhoven University of Technology, group of Prof. Grisoni
  • 2023.02 - 2023.06

    Berlin, Germany

    Researcher Intern
    Microsoft Research, AI for Science
  • 2019.09 - Present

    Berlin, Germany

    Research Fellow
    Free University of Berlin, group of Prof. Noé
    • Led a drug discovery project for TMPRSS2 inhibition which identified a novel nanomolar inhibitor and a combination preparation for treating COVID-19.
    • Majorly contributed to implementing a virtual screening pipeline comprising preparation (Schrödinger, MGLTools), docking (Smina), MD simulations (OpenMM), custom scoring (MDTraj) and active learning on molecular representations (Scikit-learn).
    • Explored different graph neural network architectures. Developed a hierarchical model (PyTorch Geometric) for protein structure representation learning.
  • 2019.01 - 2019.05

    Stockholm, Sweden

    Trainee
    Science for Life Laboratory, group of Prof. Elofsson
    • Developed a deep learning method (Keras, Biopython) based on three-dimensional convolutional neural networks for protein model quality assessment.
  • 2017.05 - 2017.03

    Utrecht, Netherlands

    Trainee
    Utrecht University, group of Prof. Bonvin
    • Analyzed intermolecular contacts/energetics and trained a random forest model (Scikit-learn, Biopython) to distinguish biological from crystallographic interfaces.

Volunteer

Skills

Programming/scripting
Python
Bash
R
MATLAB
Java
C/C++
Machine learning
PyTorch
PyTorch Geometric
Scikit-learn
Tensorflow
Keras
Data science
NumPy
SciPy
Pandas
Matplotlib
Computational chemistry
RDKit
Open Babel
Autodock Vina
Biotite
Biopython
MoleculeKit
PDBFixer
MDTraj
OpenMM
PyEMMA
Other
Git
GitHub Actions
LaTeX
Slurm
Docker

Languages

Montenegrin
Native
English, Italian
Full proficiency
German, Spanish
Elementary proficiency

Certificates

IELTS Academic - level C1
British Council 2017-06

Awards

Interests

Sports and outdoor activities
Padel
Running
Swimming
Scuba diving
Hiking
Creative activities
Painting
Pottery
DIY and crafting
Travel video editing
Music
Alternative rock
Blues, R'n'B and soul
Cinematic and orchestral
Books
Popular science
Absurdist fiction
Mystery fiction