Freelancer profile translated to English.

Description

I help life science teams with RStudio & Jupyter to get decision-ready analyses: tidyverse/Quarto reports and clear effect sizes with CIs. For image data, I set up ImageJ/FIJI or Python pipelines (segmentation → feature table) and transfer them directly into statistics. Upon request, PyTorch/TensorFlow for lightweight DL baselines (classification/regression) – reproducible with renv/Conda and 1-command builds. Specialized in Image-to-Features and methodologically sound reports (peer-review proven). Send me your data type + goal – I'll reply with a lean scope & timeline.

Languages

German
Native or bilingual
English
Fluent
Spanish
Conversational
French
Basic

Workplace preferences

Can work on-site

Frankfurt am Main (up to 50km)

Institut für klinische Pharmakologie
Data Scientist
April 2022 - February 2024 (1 year and 10 months)
Building data & ML workflows: End-to-end pipelines in R/Python (RStudio/Jupyter, Linux/Conda, partly GPU) for clinical datasets; feature engineering, model selection (classical & ML).

Methodological contributions to clinical pharmacology: Review and methodology work on AI/ML in clinical pharmacology; elaboration of application areas, opportunities, and limitations.

Meta-/Scientometric evaluation of AI in pain research: Data-driven mapping of the field and trends.

Publication-ready preparation: Manuscript creation, figures/graphics, and transparent methods reports for peer review.
Universität Trier
Research Assistant
December 2011 - June 2012 (6 months)
Operationally supported in vitro quantification of PPD derivatives by RP-HPLC – from sample preparation through measurement execution to an evaluable data basis
Freelancer
Freelance Data Scientist
September 2025 - Today (9 months)
RStudio & Jupyter workflows for statistics and quantitative image analysis. In R: tidyverse, Quarto/R Markdown, reproducible pipelines, contrasts, effect sizes with CIs, clean diagnostics. In Python: Jupyter notebooks with PyTorch/TensorFlow (classical models to lightweight DL: classification/regression, simple CNN pipelines), metrics & visualizations. I process image data with FIJI/ImageJ or Python and transfer it directly to statistical evaluation. Delivery as a runnable repo (RStudio/Jupyter) with 1-command build, README, and short decision report.
Jupyter Notebook Python ImageJ