Research Scientist – Foundation Models and Multimodal AI

Owkin | Artificial Intelligence / Biotechnology

2h 16m ago

Remote In-Country

Full Time

France, Germany, United Kingdom

About us

Owkin is an agentic AI company pioneering Biological Artificial Superintelligence to solve problems in biology where human researchers alone have failed.
Owkin builds K Pro - an AI scientist for pharmaceutical research and strategic decision-making. K Pro orchestrates a suite of AI skills and tools to decode complex biology, accelerate research, and dramatically increase productivity.
K Pro is built on Owkin’s unrivalled multimodal patient data network, state-of-the-art AI for biology and a decade of experience working with pharmaceutical partners.

Position is based in our Paris office or remotely in France, UK, Germany.

Please submit your CV in English

About the role:

We are seeking a driven Research Scientist with hands-on expertise in Foundation Models (FMs) and Multimodal AI to help drive the next phase of our machine learning platform.

In this role, you will bridge cutting-edge AI research and practical application. Rather than training massive base models from scratch, your focus will be on fine-tuning, adapting, and aligning open-weights foundation models across diverse data modalities (e.g. structured tabular data, single-cell scRNA-seq, image representations, and natural language) to build multimodal systems.

In particular, you will be responsible for:

Model adaptation and optimization: Adapt, fine-tune, and evaluate state-of-the-art open-source foundation models (e.g., domain-specific transformers, vision-language models) using fine-tuning and transfer learning strategies.
Multimodal integration: Design and implement cross-modal alignment architectures (e.g., contrastive learning, projection heads, unified embedding spaces) to fuse structured, sequential, and text/image data layers.
Pipeline and evaluation engineering: Build rigorous, domain-tailored benchmarking frameworks and evaluation pipelines to validate model reliability, generalization, and downstream task performance.
Applied R&D collaboration: Work closely with cross-functional teams (data engineers, domain specialists, and software engineers) to integrate adapted models into scalable production and inference workflows.
Low-hanging fruit innovation: Identify high-impact, incremental improvements in data pre-processing, embedding representation, and fine-tuning to maximize model performance without unnecessary computational overhead.

About you

Required qualifications / experience:

Education: Ph.D. or Master’s degree in Computer Science, Machine Learning, Mathematics or a related quantitative field.
Experience: 2-4 years of post-degree experience applying deep learning and foundation models to complex datasets.
Foundation models: Demonstrated experience adapting and fine-tuning foundation models.
Multimodal learning: Proven experience with cross-modal representation learning, joint embedding alignment (e.g., CLIP-style architectures), or vision-language models.
Technical stack:
- Advanced proficiency in Python and modern deep learning frameworks (PyTorch, Hugging Face ecosystem).
- Hands-on experience with fine-tuning techniques.
- Familiarity with data engineering tools for handling large-scale or multi-modal datasets (Parquet, AnnData, HDF5).

Preferred qualifications/bonus:

Experience working with biological or omics data (e.g., scRNA-seq, spatial transcriptomics, clinical and/or histology data).
Experience deploying or scaling models using containerized environments (Docker, Kubernetes) and cloud infrastructure (AWS/GCP).
Track record of open-source contributions or peer-reviewed publications at top ML conferences (NeurIPS, ICML, ICLR, CVPR) or domain-specific journals.

#LI-HB1

What we offer

Flexible work organization
Friendly and informal working environment
Opportunity to work with an international team with high technical and scientific backgrounds

Recruitment Process & Security

Please complete the form and submit your CV.
Owkin is an Equal Opportunity Employer. All qualified applicants will receive consideration for employment without regard to race, sex, gender, sexual orientation, age, color, religion, national origin, protected veteran status or on the basis of disability.
Owkin is a great place to work. As a coveted workplace we are, unfortunately, vulnerable to recruitment phishing scams. We urge all job seekers and candidates to be wary of potential scams. Most of these have individuals posing as representatives of prominent companies, including Owkin, with the aim of obtaining personal, sensitive, or financial information from applicants. These scams prey upon an individual’s desire to obtain a job and can sometimes “feel” like a genuine recruitment process. Some red flags are identified below. Should you encounter a recruitment process that claims to be for Owkin but is not consistent with the below, please do not provide any personal or financial information:
Legitimate Owkin recruitment processes include communication with candidates through recognized professional networks, such as LinkedIn.
Communication is always through an official Owkin email address (from the @owkin.com domain), over the phone or through our applicant tracking system (Ashby).
The Owkin talent team do use platforms such as LinkedIn, however if you have any concern or doubt about this contact, please ask for them to send an email from @Owkin.com.
The Owkin talent team will not solicit personal data from candidates during the application phase including, but not limited to, date of birth, social security numbers, or bank account information;
Legitimate Owkin interviews may be conducted over the phone, in person, or via an approved enterprise videoconferencing service (Google Meets). They will not occur via Signal, Telegram or Messenger
Owkin offers of employment are based on merit and only extended once a candidate has interviewed with members of the talent and hiring team. Offers will be extended both verbally and in written format.

If you think that you have been a victim of fraud,

Check the identity of the talent team on LinkedIn
Check our senior team on our website
Check the existence of the position on our website:
contact the following authorities:
- [FR]
- [UK]
- [US]

Required Skills

pythonpytorchhugging_face

Required Languages

🇬🇧 English

About us

Position is based in our Paris office or remotely in France, UK, Germany.

Please submit your CV in English

About the role:

We are seeking a driven Research Scientist with hands-on expertise in Foundation Models (FMs) and Multimodal AI to help drive the next phase of our machine learning platform.

In particular, you will be responsible for:

Model adaptation and optimization: Adapt, fine-tune, and evaluate state-of-the-art open-source foundation models (e.g., domain-specific transformers, vision-language models) using fine-tuning and transfer learning strategies.

Multimodal integration: Design and implement cross-modal alignment architectures (e.g., contrastive learning, projection heads, unified embedding spaces) to fuse structured, sequential, and text/image data layers.

Pipeline and evaluation engineering: Build rigorous, domain-tailored benchmarking frameworks and evaluation pipelines to validate model reliability, generalization, and downstream task performance.

Applied R&D collaboration: Work closely with cross-functional teams (data engineers, domain specialists, and software engineers) to integrate adapted models into scalable production and inference workflows.

Low-hanging fruit innovation: Identify high-impact, incremental improvements in data pre-processing, embedding representation, and fine-tuning to maximize model performance without unnecessary computational overhead.

About you

Required qualifications / experience:

Education: Ph.D. or Master’s degree in Computer Science, Machine Learning, Mathematics or a related quantitative field.

Experience: 2-4 years of post-degree experience applying deep learning and foundation models to complex datasets.

Foundation models: Demonstrated experience adapting and fine-tuning foundation models.

Multimodal learning: Proven experience with cross-modal representation learning, joint embedding alignment (e.g., CLIP-style architectures), or vision-language models.

Technical stack:

Advanced proficiency in Python and modern deep learning frameworks (PyTorch, Hugging Face ecosystem).
Hands-on experience with fine-tuning techniques.
Familiarity with data engineering tools for handling large-scale or multi-modal datasets (Parquet, AnnData, HDF5).

Preferred qualifications/bonus:

Experience working with biological or omics data (e.g., scRNA-seq, spatial transcriptomics, clinical and/or histology data).

Experience deploying or scaling models using containerized environments (Docker, Kubernetes) and cloud infrastructure (AWS/GCP).

Track record of open-source contributions or peer-reviewed publications at top ML conferences (NeurIPS, ICML, ICLR, CVPR) or domain-specific journals.

#LI-HB1

Recruitment Process & Security

Please complete the form and submit your CV.

Owkin is an Equal Opportunity Employer. All qualified applicants will receive consideration for employment without regard to race, sex, gender, sexual orientation, age, color, religion, national origin, protected veteran status or on the basis of disability.

Owkin is a great place to work. As a coveted workplace we are, unfortunately, vulnerable to recruitment phishing scams. We urge all job seekers and candidates to be wary of potential scams. Most of these have individuals posing as representatives of prominent companies, including Owkin, with the aim of obtaining personal, sensitive, or financial information from applicants. These scams prey upon an individual’s desire to obtain a job and can sometimes “feel” like a genuine recruitment process. Some red flags are identified below. Should you encounter a recruitment process that claims to be for Owkin but is not consistent with the below, please do not provide any personal or financial information:

Legitimate Owkin recruitment processes include communication with candidates through recognized professional networks, such as LinkedIn.

Communication is always through an official Owkin email address (from the @owkin.com domain), over the phone or through our applicant tracking system (Ashby).

The Owkin talent team do use platforms such as LinkedIn, however if you have any concern or doubt about this contact, please ask for them to send an email from @Owkin.com.

The Owkin talent team will not solicit personal data from candidates during the application phase including, but not limited to, date of birth, social security numbers, or bank account information;

Legitimate Owkin interviews may be conducted over the phone, in person, or via an approved enterprise videoconferencing service (Google Meets). They will not occur via Signal, Telegram or Messenger

Owkin offers of employment are based on merit and only extended once a candidate has interviewed with members of the talent and hiring team. Offers will be extended both verbally and in written format.

If you think that you have been a victim of fraud,

Check the identity of the talent team on LinkedIn

Check our senior team on our website

Check the existence of the position on our website:

contact the following authorities:

[FR]
[UK]
[US]