Machine Learning Researcher

Postdoctoral Position in Multimodal & Generative AI for Biomedical Imaging and Genetics

For the Italian Institute of Technology (IIT), we are looking for a PhD holder with advanced expertise in Machine Learning and Computer Vision to join the Artificial Intelligence for Good (AIGO) research group and work on multimodal and generative models for biomedical imaging, with the goal of predicting genetic information and clinical risk from medical images.

Keep reading if you:

• Hold a PhD in Artificial Intelligence, Machine Learning, Computer Vision, Computer Science, Engineering, Physics, Mathematics, or a related field
• Have a strong research background in Machine Learning, Deep Learning and Computer Vision, preferably applied to medical/biomedical imaging or other scientific domains
• Have worked with multimodal models or approaches integrating images with structured, biological, clinical or genetic data
• Have published research in Machine Learning, Deep Learning and/or Computer Vision journals and conferences

This position may not be for you if you:

• Are looking for a permanent position
• Come from a genetics or genomics background but have never worked with Deep Learning models
• Have mainly used existing Machine Learning or Deep Learning models as ready-made tools, without being directly involved in their development, training or adaptation to complex scientific problems

Who you will be working with

The selected candidate will join the Italian Institute of Technology (IIT), a research institution established to promote technological development and advanced scientific training in Italy.

The position is within the Artificial Intelligence for Good (AIGO) research group, led by Prof. Vittorio Murino, an internationally recognised researcher in Artificial Intelligence.

The group currently includes approximately 25 PhD students, postdoctoral researchers and research scientists. Its research focuses on learning from imperfect data, particularly in multimodal settings, including unsupervised, semi-supervised and self-supervised learning, as well as learning from weakly labelled, noisy, imbalanced or biased data. Other research areas include domain adaptation and generalisation, few-shot and zero-shot learning, continual learning and learning from biased data.

The group also works on generative models, with particular attention to recent multimodal foundation models, including Large Language Models (LLMs) and Vision-Language Models (VLMs). Further research focuses on lightweight Machine Learning approaches aimed at developing energy-efficient AI technologies, including applications on edge devices such as robots.

AIGO also develops AI methods that incorporate ethical considerations, privacy, fairness and robustness from the ground up, with the goal of developing Deep Learning techniques that are explainable, reliable and transparent. Its main application areas include biomedicine, biology, neuroscience and healthcare.

AIGO collaborates with several international universities and research centres, including close collaborations with the Universities of Genoa and Verona.

The project

The position is funded by Dompé Farmaceutici within the project “AI Driven Prediction of Glaucoma Linked Genetic Variants from OCT Retinal Imaging.”

The project aims to explore whether retinal OCT images and heterogeneous biomedical data can be used to predict genetic factors and the risk of developing glaucoma.

The research will focus on developing models capable of learning phenotypic representations from images that can serve as predictive indicators of underlying genotypic factors.

The project therefore aims to move beyond traditional approaches based primarily on statistical association analysis, leveraging the ability of Deep Learning to learn from high-dimensional imaging data together with contextual and multimodal information.

What you will work on

You will not simply apply existing models to a biomedical dataset. Instead, you will contribute, as part of the research team, to the design, development and validation of new methodological solutions.

In collaboration with the AIGO team, you will:

• Design and investigate Machine Learning and Deep Learning models capable of linking medical imaging data to biological, genotypic or genetic risk factors
• Work on multimodal and generative models, representation learning and data-driven approaches applied to images and heterogeneous biomedical data
• Address challenges related to limited supervision, data heterogeneity and the availability of multiple sources of information
• Conduct research both independently and in collaboration with other AIGO researchers
• Supervis