Build the science that shapes the future of human health.
Application closing date: 30.04.2026

Join a place where ambitious science thrives
Human Technopole (Milan) is a rapidly expanding life science institute where international researchers and cutting-edge technologies converge to accelerate biomedical discovery. Our mission is to transform bold scientific ideas into advances that improve human health.

Within this mission, the Jug Group develops advanced computational methods and open scientific software to extract knowledge from complex biological imaging data. We work at the intersection of artificial intelligence, microscopy, and quantitative biology, translating cutting-edge research into robust tools that are used by researchers worldwide. Our work contributes directly to enabling reproducible science, scalable data analysis, and next-generation experimental workflows that link data acquisition, analysis, and interpretation.

We are seeking an ambitious Postdoc to join us in shaping how AI-enabled methods and research software empower modern biomedical research.

Your mission

As a Postdoc you will help build the next generation of multimodal and explainable AI models for personalized risk prediction. You will develop methods to integrate diverse medical imaging modalities (e.g., X-rays, MRI, Ultrasound) with Electronic Health Record (EHR) data across heterogeneous cohorts and healthcare systems, including EHR data curation and extraction at scale.

The project is part of the Scientific AI Flagship at Human Technopole, and you will be embedded within the Jug Group. In this role, you will work in a highly multidisciplinary environment, collaborating closely with radiologists, software engineers, and data scientists to advance clinically relevant AI methods. Your research is expected to lead to high-impact publications, tackling methodology for synthetic patient alignment and iterative self-improvement for robust generalization, explainability, and uncertainty assessment. Where appropriate, outcomes may also support intellectual property development and the translation of methods into deployable solutions.

Grow your skills
You will enhance your scientific and professional skills by:

• Designing and analysing multimodal deep learning models for time-specific cancer risk and time-to-event prediction by integrating imaging with longitudinal Electronic Health Record (EHR) signals.
• Building scalable pipelines for EHR data curation and extraction (structured codes, labs, medications, etc.), including the handling of missing data and irregular sampling.
• Developing methods for synthetic patient alignment and iterative self-improvement, leveraging non-matched cohorts to harmonize modalities and reduce dataset, device, and demographics dependences.
• Implementing robust training strategies for heterogeneous clinical data (e.g., weak supervision, semi/self-supervision, domain generalization, calibration-aware objectives) with explicit mitigation of confounding, selection bias, leakage, and study-design effects.
• Creating clinically meaningful explainability for multimodal predictors (beyond heatmaps), including reasoning summaries, counterfactual/what-if analyses, causality, and uncertainty-aware explanations tailored to radiology workflows.
• Developing software tools that supports reliable model comparison, ablations, and useful documentation (for human users and agentic use).
• Communicating results effectively through manuscripts, talks, and open-source releases, contributing to high-impact publications and reusable methods for trustworthy biomedical AI applications.
• Collaborating closely with other scientists, radiologists, clinicians, and engineers.

Human Technopole supports career development through training, mentoring and dedicated learning opportunities.

What you'll bring

Essential

• PhD Degree in a relevant scientific field (e.g. computer science, data science, mathematics, statistics, engineering, physics, or related).
• Provable deep learning track record and practical expertise (e.g. with VAEs, GANs, diffusion models, transformers, representation learning, contrastive methods).
• Solid experience with Python and Pytorch and good command of software development tools such as GIT.
• A good track record of authoring scientific publications in a relevant field.
• Fluency in English.

Preferred

• Knowledge of statistics for clinical prediction (confounding, selection bias, information leakage, study design impact like case-control vs cohort, etc.).
• Strong grasp of statistical/ epidemiological principles e.g. risk prediction and survival/time-to-event modelling.
• Experience with uncertainty aware evaluation: confidence intervals/bootstrapping, hypothesis testing, subgroup/stratified performance analysis.
• Large cohort and datasets handling.
• Experience with Multimodal Deep Learning approaches, representation learning, etc.
• Experience with Electronic Health Record data (structured codes (ICD), labs, meds, notes and/or temporal event sequences; irregular sampling).
• Experience with medical imaging modalities such as mammography/ DBT, x-rays, MRI, ultrasound, etc.

Organisational and social skills

• Ability to work accurately and self-motivated with a high level of attention to detail.
• Ability to prioritize and coordinate multiple workstreams (research, software development, data management) while meeting internal milestones and external deadlines.
• Excellent communication skills (spoken and written) and ability to collaborate across disciplines.
• Collaborative mindset and proven ability to work effectively in interdisciplinary teams.

At HT, your discoveries contribute to a global effort to improve human health.

Why Human Technopole
HT offers an international and dynamic workplace, competitive welfare provisions, flexible working policies and relocation support. Researchers moving to Italy may benefit from attractive tax benefits. We promote work–life balance and provide parental support initiatives.

How to apply
Submit:

• CV.
• Motivation letter (English).
• Contact details of two referees.

This is a 3-year contract offered under CCNL Chimico Farmaceutico, Level B2

Salary: up to € 43.000,00, depending on the seniority of the candidate.
The position is based in Milan, Italy, within our vibrant international campus.
We strongly encourage applications from candidates belonging to protected categories (L. 68/99).