Research Assistant (f/d/m)

The Peter L. Reichertz Institute for Medical Informatics () of TU Braunschweig and MHH offers a full-time position as Research Assistant in expert- and AI-supported federated learning of clinical decision support models to be filled from November 1st, 4.

PLRI, one of the largest university-based centers for medical informatics in Germany, belonging to two leading universities, comprising an institute of technology and a medical school, located in one of Europe's most research-intensive regions, offers excellent research opportunities.

We collaborate in interdisciplinary projects with health care centers, research institutes, enterprises and public organizations in order to shape the future of healthcare and medicine.

Our activities range from the local level as with the institutions in Braunschweig and Hannover, through regional, national and global corporations.

The advertised full-time position is part of the interdisciplinary, cross-site third-party funded research project "PRETTY - Personalized PREdiction of Transplant ToxicitY through federated learning from data, expert opinions and patient perspectives".

In PRETTY, we aim at helping leukemia patients who may suffer from nephrotoxicity as a severe side effect of a allogeneic hematopoietic cell transplantation (alloHCT) treatment.

Very little is known about the risk factors that favor nephrotoxicity in an individual patient. The main obstacles to gain further knowledge are the lack of comprehensive data collection and analysis of large patient cohorts.

We will address these obstacles by enabling prospective data integration in four large German university hospitals, establishing four local personalized prediction models for nephrotoxicity in alloHCT treatment, and integrating the local models into a unified federated prediction model.

We hypothesize that the federated prediction model yields an improved accuracy over the non-federated models. Unlike most federated learning approaches, which focus on data only, we will also integrate local expertise of those treating (physician perspective) and those being treated (patient perspective, patient reported outcomes wrt.

severity of nephrotoxicity effects) into the model learning. We expect that our approach will improve quality of life and survival of patients undergoing alloHCT, identify modifiable risk factors for nephrotoxicity, reduce hospitalizations and hospital costs, and benefit other cancer patients by model transfer.

The advertised position will be located on the MHH campus and will work closely with the hematology department at MHH as well as with the hematology departments and the data integration centers at the other four clinical sites.

International applicants will need to complete a visa process before hiring can take place.

Your task

Your main task is to develop methods for the integration of both, the physician and the patient perspective into the model learning as well as into the model evaluation process.

This comprises hybrid methods from machine learning, which can integrate (potentially conflicting) expert opinions as well as methods for model visualization and collaborative model inspection and editing.

A local team of a physician and a patient representative will support the model learning and evaluation at each of the four participating clinical sites in PRETTY.

A strong collaboration with these teams in method development as well as in evaluation is required.

You will be expected to develop innovative new approaches, to publish them in high-quality scientific publications at renowned conferences and in journals and to participate in the acquisition of follow-up third-party funds as well as in administrative tasks of the institute.

Our requirements

• obtained a Master’s degree in the area of computer sciences, data science, information systems, medical informatics, bioinformatics, business informatics, mathematics, or a comparable course of study
• expertise in the development of AI algorithms and model learning from complex data experience in federated learning, (clinical) decision support, and / or Bayesian networks is beneficial, and visualization of complex data and / or user interface design
• good programming (Python, R, etc.) and software engineering skills in the analysis of complex (research) data
• willingness to work closely with an interdisciplinary project team
• good oral and writing skills in English skills in German are beneficial
• self-initiative and result-oriented working style

We offer

• a full-time position with .5 hours per week, initially limited until October st 6, with one of the largest employers in the state of Lower Saxony
• a salary in accordance with TV-L with the benefits of the public sector (e.g. company pension scheme and additional insurance through VBL)
• a stimulating position at the interface of research, care and science management
• a collegial induction programme in a motivated, team-oriented, international and interdisciplinary team
• flexible working hours within the framework of official agreements and the possibility of mobile working
• the possibility of obtaining a PhD
• personal and professional development opportunities - supported by our wide range of internal and external training and development programmes
• an excellent public transport connections and a job ticket for local public transport
• an excellent range of sports, counselling and prevention programmes - because your health is important to us
Vor 7 Tagen