Research & Development

We dedicate a significant share of our budget to research and publication. Our work focuses on artificial intelligence applied to healthcare: research, benchmarks, evaluation and security.

R&D

Research

AI applied to healthcare

Benchmarks

Reproducible evaluation

Security

Robustness & safety

Research Areas

Applied research on machine learning and AI methods for laboratory and clinical data, from problem framing to published results.

AI Research

Healthcare applications

Benchmarks and datasets built from real laboratory and clinical tasks, so models are compared on what matters in practice.

AI Benchmarks

Realistic tasks

Independent evaluation of AI models and products: performance, calibration, bias across populations and behaviour on data that differs from training.

AI Evaluation

Independent assessment

Security testing of AI systems used in healthcare: adversarial robustness, data leakage, prompt injection and misuse of language models.

AI Security

Attacks & defences

We publish our methods and results and share them with the scientific and healthcare communities.

Publications

Sharing results

Open to joint projects with laboratories, hospitals, academic teams and companies, from research questions to proofs of concept.

Collaboration

Joint projects

Work with us on research

A model to evaluate, a benchmark to build or a research question to explore? Let’s discuss it.