PhD Students (f/m/d) at the Helmholtz School for Integrated Data Science in Environmental and Life Sciences (IDEAS)

The Helmholtz School for Integrated Data Science in Environmental and Life Sciences (IDEAS) connects the domain-science expertise of UFZ and HZDR with the data/information science strength of Leipzig University (LU) and Dresden University of Technology (TUD), supported by CASUS as an interdisciplinary bridge. IDEAS is part of the Helmholtz Information & Data Science Schools under the Helmholtz Data Science Academy (HIDA).

Our research focus
IDEAS advances and applies modern data science to complex challenges in environmental and life sciences (e.g., machine learning, explainable AI, uncertainty quantification, and AI-ready FAIR data and research data management).

What you can expect at IDEAS
IDEAS offers structured, interdisciplinary supervision and training, including joint supervision across disciplines, a Thesis Advisory Committee (TAC), a tailored curriculum, and cohort activities (seminars, hackathons, retreats), plus strong career development and networking through the IDEAS/HIDA ecosystem.

PhD topics

This collective call includes 8 PhD topics, of which 6 positions will be funded. Applicants can be considered for multiple projects and will be matched through a structured selection and ranking process.

• Climate Disasters

Climate disasters cause major human and economic losses, but it remains difficult to explain why impacts differ across places and time. This PhD project combines newly available disaster, socio-economic, and satellite datasets with interpretable machine learning to disentangle the roles of hazard intensity, exposure, vulnerability, and environmental conditions. You will develop data-science methods that address biased impact records and the spatio-temporal structure of the data, with the goal of improving our understanding of what drives disaster impacts and making climate-risk assessment more reliable.

• Beyond Traditional Monitoring

Traditional chemical monitoring can miss short-lived or poorly captured pollution events—yet these may be visible in newspapers, local reports of fish kills, or social-media complaints about odours and discoloured rivers. This PhD project develops AI-based methods to integrate such event signals with regulatory intelligence and chemical data in the SARDINE platform. The aim is to strengthen mixture risk assessment and improve freshwater protection by revealing monitoring blind spots and better characterising real-world exposure.

• Decoding Protein Darkmatter

Many proteins remain "functionally uncharacterised,” representing a major blind spot in biology and biotechnology. In this PhD project, you will use protein language models trained on billions of sequences to study how scale, and evolutionary and ecological diversity, shape model generalisation. You will develop interpretable and robust embeddings and validate predictions with Helmholtz lab partners, ultimately enabling discovery of novel enzymes relevant for sustainable biotechnology.

• Estimating LLM Biodiversity

Large language models rely heavily on internalised factual knowledge, but the true extent of that knowledge is hard to measure. This PhD project reframes LLM knowledge as "knowledge diversity,” drawing an analogy to biodiversity in ecology, and applies ecological estimation methods to infer the amount of knowledge from limited samples. By bridging computer science and statistical ecology, the project aims to produce reliable knowledge estimates for LLMs while also stress-testing and advancing biodiversity estimators at scale.

• AI Against Cancer

This PhD project develops multimodal AI that integrates clinical text and medical imaging to reduce overdiagnosis, overtreatment, and unnecessary monitoring in prostate cancer—supporting better decisions for thousands of patients. You will combine foundation models and LLM-based agents with large-scale computing in an international team spanning Helmholtz (Germany) and Danish clinical and technical partners, with close day-to-day clinical supervision. The goal is clinically relevant decision support with direct, measurable patient impact.

• Inequalities in Climate Discourse

Political attention to climate change and disasters varies across countries and societal groups, and these differences may shape climate action. This PhD project uses large-scale text analysis, NLP, and machine learning to quantify and explain variation in climate-related attention across UN speeches and parliamentary debates. You will link discourse patterns to policy and action indicators, helping to clarify how representation and political salience influence climate action.

• EXACT

In this PhD project, you will develop explainable and uncertainty-aware graph-based AI models to predict chemical toxicity for trustworthy environmental and regulatory decision-making. Working at the interface of data science, graph theory, and computational toxicology, you will integrate large public toxicity datasets and translate model outputs into chemically interpretable insights. The project aims for methods that are both accurate and usable in practice.

• TrustSeg

Medical image segmentation is central to radiotherapy and surgery, but clinical deployment requires not only accuracy—it requires knowing when a model is uncertain. This PhD project develops uncertainty-aware segmentation methods that quantify and communicate confidence to support safer decision-making in safety-critical settings. You will work closely with clinicians and researchers to build trustworthy computer-vision systems designed for real-world clinical use.

All projects are described in detail (including your tasks, your profile and the application documents you need to submit) on our IDEAS website.

Your tasks

Your tasks will depend on the project you are matched to.

Across all projects, you will conduct original PhD research at the interface of data science and domain science, contribute to publications and scientific dissemination, and participate in IDEAS training and cohort activities (e.g., seminars, coursework, and community events).

Your profile

Requirements vary by project, but what generally applies across the call:

• A very good Master’s degree (MSc or equivalent) in a relevant field (e.g., data science, computer science, mathematics/statistics, physics, environmental sciences, life sciences/bioinformatics, computational social science, or related areas), depending on the project.
• Strong programming / data analysis skills and motivation to work with large, complex datasets and modern ML/AI methods.
• A strong interest in interdisciplinary research bridging data science and application domains, and the ability to collaborate in diverse teams across institutions and locations.
• Very good English skills (written and spoken) for work in an international research environment.

Application + Hardfacts

• Place of Work

Leipzig or Dresden, depending on the project; mobile work possible

• Working time

100% (39 h/week)

• Contract limitations

Limited contract / 3 years (extension by a fourth year is possible)

• Salary

Remuneration according to the TVöD piblic sector up to pay grade 13 including attractive public-sector social security benefits

• Contact

Sandra Hille (UFZ - Tel. +49 341 6025 4674)

Anne Pidt (HZDR - Tel. +49 351 260 4716)

Please submit your application via IDEAS website until February 22, 2026.