Belle.ai is looking for an Applied Mathematics & AI Intern to join our AI team and work on real-world problems at the intersection of medical imaging, computer vision, applied mathematics, and data analysis.
You will work with dermoscopic images and medical data alongside AI engineers and medical experts, with significant freedom to experiment and propose new approaches.
The Role
This internship covers two complementary work streams within our AI team.
Dermoscopic image analysis
- Develop and improve image-processing and computer-vision pipelines for dermoscopic images, using mathematical methods and machine learning for segmentation and quantitative image analysis.
- Experiment with and benchmark new approaches.
- Integrate AI models into APIs, using tools such as Docker.
Model evaluation
- Build a reusable evaluation framework - quantitative metrics, qualitative failure analysis, and clear documentation
- Design metrics and decision thresholds under asymmetric clinical risk and assess robustness across image quality and patient subgroups.
- Investigate failure cases qualitatively: look at what the model gets wrong and explain why, not just report the number.
Across both
- Analyze experimental results and structured datasets using statistics and data analysis to identify relevant insights.
- Share and communicate results, present findings, and contribute to benchmarking and validation.The two workstreams are distinct, and you will move between them depending on project needs. Together they give you experience across both sides of medical AI - building models and rigorously validating them.
Who We’re Looking For
Technical Background
- Good background in applied mathematics, data science, computer vision, or a related quantitative field.
- Good Python skills and experience with scientific and AI libraries such as NumPy, pandas, and PyTorch.
- Good understanding of statistics, data analysis, machine learning, and deep learning.
- Familiarity with model evaluation metrics - confusion matrices, precision/recall, top-k accuracy, or metrics for segmentation and detection - and an understanding of what each does and doesn't tell you.
- Ability to turn mathematical ideas into practical, testable solutions and work with imperfect real-world data.
- Professional English required.
Real projects matter to us. Personal projects, GitHub repositories, competitions, research projects, or other side projects are a strong plus.
Mindset
We are looking for someone who is:
- Curious and eager to learn - you are willing to explore new technical and scientific fields.
- Open-minded and adaptable - you are comfortable working in a multicultural and multidisciplinary environment.
- Proactive and rigorous - you experiment, challenge assumptions, and take ownership of your work.
The candidate should be open to occasional international travel for collaboration, projects, or other professional opportunities.
Why Belle.ai
- Work on real medical imaging problems with direct product applications.
- Collaborate with AI engineers, dermatologists, and medical experts in a strong team-oriented environment where exchange, sharing progress, and collective problem-solving are highly valued.
- Develop skills across applied mathematics, computer vision, AI, and statistics.
- Work in an international environment across France, Vietnam, and the US.
- Have the freedom to experiment, propose ideas, and take ownership of technical problems.
How to Apply
Send us your CV and, if available, something that shows what you can build: GitHub, or portfolio to [email protected] / Subject: AI Scientist intern| Full Name
We are particularly interested in candidates who can show us a project they are proud of and explain the problem, their approach, and the results.
Rémunération : à partir de 800,00€ par mois
Lieu du poste : En présentiel
