About the role:

This position is based in Paris, France. Ideally you are able to join the team 3rd March 2025. Please submit your CV in English.

At Owkin, we develop proprietary models designed to improve patient stratification by accurately predicting clinical endpoints. These models are central in our projects and integrate diverse data sources, including molecular profiles, histological imaging, and clinical information, to generate a comprehensive understanding of patient characteristics.

The goal of this internship is to enhance our ability to extract biological insights and meaningful patient representations from MOSAIC, our unique and cutting-edge multimodal oncology database. This work will help us to distill complex multimodal patient subtypes into clinically actionable biomarkers with a specific focus on Antibody Drug Conjugates (ADCs).  In this context, you will focus on advancing translational modeling and signal extraction techniques. This includes tasks such as feature engineering, feature selection, model training and evaluation, knowledge distillation and signal quantification within high-dimensional, multimodal datasets.

You will join a dynamic and innovative environment, working on projects that bridge complex molecular insights with real-world clinical applications. Your work will be integrated in Owkin’s technical stack, and will directly contribute to improving Owkin’s internal capabilities in our journey to discover new targets, subtype patients more accurately, and eventually identify the right treatment for the right patient.

In particular, you will:

As your main missions, you will:

• Collaborate closely with, and receive mentorship from, the other members of the R&Tech team.
• Engineer new and complex features from multimodal MOSAIC data to help guide patient selection for ADCs.
• Work closely with clinical and pathology experts to assess the robustness and validity of features extracted from MOSAIC data.
• Develop and benchmark models that enrich likely ADC responders.
• Identify clinically-actionable biomarkers that allow for the translation of complex patient representations and subtypes.
• Adhere to Owkin’s code quality standards, and be responsible for the re-usability of your implementations.
• Develop code that will directly enhance the capabilities of Owkin’s Discovery Engine.
• Share your findings through regular research reviews and detailed reports to the team.

About you

Required qualifications / experience:

• You are enrolled in a master’s degree in biomedical engineering, bioinformatics, computational biology, computer science, mathematics, statistics or a related field
• Authorization to work legally in France
• Fluent in English (spoken and written)
• Proficient in Python and experience with ML / DL frameworks (e.g., scikit-learn, Pytorch)
• Strong communication skills
• Good team player

Preferred qualifications/bonus:

• Experience (previous internship, project, kaggle, etc.) in ML for healthcare applications

Please submit your CV in English