since 06/2024
Independent Consultant in Computational Sciences, Data and AI
Vienna
Partnered with individuals, startups, and large organizations to design and implement bespoke computational systems tailored to their needs. Projects included:
- Built European Pharmacopoeia-compliant software at Minerva Biolabs that replaced Excel: from raw data to the reported result and its uncertainty, readable by lab staff.
- Author of synfit, open-source dose-response and drug-synergy fitting package (on PyPI), and web app (synfit.app): Hill models, explicit noise models, Bliss / HSA / Loewe / ZIP synergy.
- Author of bioio-vsi, an open-source reader for Olympus/Evident cellSens VSI microscopy images without Java (on PyPI): undocumented format reverse-engineered from sample files, validated against Bio-Formats.
- Author of wfdisc, an open-source Rust library for reading seismological waveform data (CSS 3.0 / NNSA KB Core), cross-validated against ObsPy.
- Consulted for Discovery Evolution, an in-silico biologics-discovery start-up, in its early stage.
- Developed interactive applications enabling expert users to design and explore complex sensor configurations and large-scale datasets in real time.
- Designed a procedural-support platform for field operators at an international organization in a strictly regulated environment: formalized SOPs to reduce cognitive load and error rates.
- Advised startups and research groups on data strategy, computational architecture, and analytical robustness, helping them avoid unproductive directions through early feasibility validation.
10/2022 – 04/2024
Senior Data Scientist
Zühlke Engineering Austria · Vienna
Bridged scientific and engineering domains to modernize computational systems and deploy AI-assisted internal tools for knowledge management and documentation analysis.
- Built an LLM-based framework analyzing 800+ internal SOPs at a multinational client, defining clear KPIs that enabled leadership to spot gaps and standardize communication across teams.
- Refactored and optimized large scientific codebases, under strict test regimes, for up to 15× faster performance and greater maintainability.
- Designed and implemented algorithms that enabled smooth interaction with terabyte-scale data on consumer hardware, boosting the efficiency of technical operators.
09/2019 – 09/2022
Research Scientist in Biomedical Data Analysis
Medical University of Vienna
Designed and implemented computational pipelines for biomedical research, transforming experimental data into reproducible, insight-driven results adopted by multiple research groups.
- Built an end-to-end pipeline for analysis of microscopy videos (stabilization, segmentation, multi-scale event detection with a Poisson noise model, review UI, publication-ready statistics); cut analysis from days of manual work to under an hour plus a short expert review (github.com/szarma/Physio_Ca).
- Designed and deployed a first-of-its-kind database of experimental data, with interactive apps enabling scientists to explore many experiments simultaneously, revealing effects the lab had long suspected but could not show.
- March 2020: in a few-day sprint with the EpiMath team, coded and fitted the SEIR model behind the expert paper the Austrian government cited when tightening its measures (epimath.at, ORF coverage).
06/2013 – 07/2019
Research Scientist in Computational Biology
Institute of Science and Technology Austria (ISTA) · Vienna
Focused on data-driven modeling and analysis in computational biology, using machine learning and Bayesian inference to reveal organizational principles of genomic and membrane systems.
- Biophysical models of gene regulation (protein–DNA binding) fitted to high-throughput data with CNNs (TensorFlow) and multi-tempered MCMC; two eLife papers.
- Built Bayesian models and visualization concepts for single-molecule quantitative imaging, revealing spatial patterns in live-cell microscopy data.
- Authored and taught introductory programming courses (Best Lecturer Award); teaching assistant for the PhD core course Data Science and Scientific Computing.
02/2011 – 05/2013
Research Scientist in Cosmology and Astroparticle Physics
Max Planck Institute for Physics · Munich
Modeling and simulation of high-dimensional dynamics, focused on numerical stability.
- Built and maintained simulation frameworks for nonlinear dynamical systems, combining analytical modeling with high-performance, reproducible scientific software in Fortran and Python.
- Developed analytical workflows to extract key statistics and physical insights from vast simulation datasets, transforming raw numerical output into interpretable results.