Technical lead of the DeepTech start-up project NeuroPin (https://neuropin.squarespace.com) proposing a deep learning software for clinical MRI segmentation through federated learning.- Product definition: business needs identification; competitor research.- Data management: locate the data sources to generate relevant models; gather, transform and clean the data.- Conduct proof of concept development: accurate segmentation, faster than 80% of competitors.- Team management: recruitment, monitoring (1 pers).- Set up collaborations with hospitals and Dr. Marco Lorenzi.

Development of unsupervised machine learning approaches on large multimodal databases (MRI, cognition, genetics, etc) to identify suitable participants for dementia clinical trials.- Data management: functional and structural MRI datasets with up to ~3,000 participants.- Methods implementation: dimensionality reduction (PLS/CCA, manifold learning), non-parametric testing.- Management/Training: 3 students.- Collaborations on transversal projects.- Was awarded two competitive grants (AUD 20,000).Advisor: Prof. Michael Breakspear

Taught undergraduate computer science, including advanced object-oriented programming (Java), operating system (C) and data structure/fundamental algorithms (C++)

Design of a computer vision algorithm, based on deep learning, dedicated to the automatic recognition of cortical sulci from MRI. The model makes 20% less errors and is 15 times faster than previous solutions. The pipeline is available in Brainvisa and used by researchers internationally.- Methods implementation: 3D convolutional neural network (UNet), patch-based segmentation.- Taught undergraduate computer science (C/C++ and Java). Advisor: Prof. Jean-François Mangin

Design and comparison of three computer vision algorithms for automatic recognition of local cortical sulci patterns, including one based on a 3D recurrent neural network.

Designed a spiking neural network with dynamic synapses for language classification (Python).

Analysed mass spectrometry signals from animals fed with GMO maize.