IHRE AUFGABEN

We are seeking a PhD student to investigate data-driven biodiversity classification in a stakeholder process, particularly how audio recordings from citizen scientists and satellite imagery can be integrated to map biodiversity at scale. You will be part of the BMBF-funded project “BiodivKI-2: Biodiversity Assessment of Habitat Types Using Machine Learning Based on Citizen Science Audio Recordings and Satellite Imagery” (Bio-O-Ton-2). Our overarching goal is to develop and test novel machine-learning approaches for combining acoustic and geospatial data, culminating in an interactive platform to support environmental authorities.

Special focus will be on the following key questions:

• How can acoustic data and satellite imagery be combined to classify biodiversity efficiently?
• Which habitat types can be reliably differentiated using this approach?
• How can end-users (e.g., conservation authorities) benefit from a large-scale, cost-effective classification system?

In this role, you will:

• Coordinate and refine data collection from multiple citizen-science platforms.
• Develop automated pipelines (Python-based) for cleaning, merging, formatting and visualizing/interacting with big acoustic and satellite datasets.
• Collaborate with AI specialists to train and validate deep-learning models for biodiversity classification.
• Participate in field campaigns to gather additional ground-truth sound recordings.
• Contribute to co-creation workshops with stakeholders (environmental authorities, NGOs) to align methods with real-world needs.
• Present your work at academic conferences and co-author peer-reviewed publications.

IHRE QUALIFIKATION

You must hold a Master’s degree in a relevant field (e.g., Geoinformatics, Geoecology, Computer Science, Biology, Biophysics, Environmental Science, Spatial Planning or others) but no PhD. German language skills are beneficial; English proficiency is expected. We welcome applicants who are analytical, creative, and eager to work in an interdisciplinary team.

Beneficial experience and qualities include:

• Interest and/or experience in biodiversity topics, soundscape ecology, or environmental monitoring
• Experience with geospatial data science, machine-learning libraries (e.g., TensorFlow, PyTorch), or data-visualization tools
• (Willingness to acquire) skills in Python (or similar, e.g. R/MATLAB) and web-based interactive mapping (e.g., Python Shiny)
• Familiarity with remote-sensing techniques and data (e.g., Sentinel, PlanetScope)
• Enthusiasm for transdisciplinary research and stakeholder engagement

• Coordinate and refine data collection from multiple citizen-science platforms.
• Develop automated pipelines (Python-based) for cleaning, merging, formatting and visualizing/interacting with big acoustic and satellite datasets.
• Collaborate with AI specialists to train and validate deep-learning models for biodiversity classification.
• Participate in field campaigns to gather additional ground-truth sound recordings.
• Contribute to co-creation workshops with stakeholders (environmental authorities, NGOs) to align methods with real-world needs.
• Present your work at academic conferences and co-author peer-reviewed publications