Data Scientist and Architect

6/2024 – 3/2025

Two-stage project focusing on (1) building a scalable data warehouse for integrating and processing heterogeneous biodiversity data, and (2) implementing advanced machine learning workflows to model species distributions under varying environmental conditions.

Design and implementation of a scalable database architecture
- Developed a DuckDB-based solution for local data storage and processing
- Designed a normalized schema to support diverse ecological data types
- Implemented spatial operations for geographic data processing

Development of automated data pipelines
- Created an R package offering a standardized interface for data ingestion, processing, and extraction
- Built automated workflows for orchestration and reproducibility using targets
- Integrated heterogeneous data sources, including species occurrences, taxonomy, and environmental raster layers

Predictive modeling using machine learning and deep learning
- Built end-to-end modeling pipeline to predict species distributions at continental scales
- Deployed different model types under shared feature processing, cross-validation, and model evaluation regime
- Developed embedding architecture to improve information sharing across species in Neural Networks

R (Programmiersprache), Datenmodelierung, SQL, Neuronale Netze, Torch, Python

5/2024 – offen

Development of a data exploration and visualization platform to support plant breeding decisions at KWS Saat. The system provides an intuitive interface to internal data sources and enables staff researchers to explore complex interactions between gene variants, crop characteristics and physiological pathways.

Development of a REST API for graph database integration
- Built a FastAPI-based service layer for interacting with a Neo4j graph database
- Designed efficient, targeted query endpoints
- Integrated authentication and access control mechanisms

Design and implementation of an interactive analytics dashboard
- Developed a web interface for interactive data exploration using Plotly Dash
- Implemented network visualizations with Cytoscape for complex relationships
- Enabled real-time data access via API integration
- Incorporated AI-assisted features through external API services

API-Entwickler, Datenbank-Analytiker, Docker, Python, Web Entwicklung

6/2023 – 6/2024

Automation and optimization of MS PowerQuery-based workflows to structure and prepare multi-channel marketing campaigns (Google, Instagram, Youtube, etc.).

Power Bi, Google Cloud, Microsoft Excel