E-Linus

• Orchestrated the storage of raw sensoristic information combined with vital signals from smartwatches of test patients in La Pace retirement home in Sutri, Latium.

• Engineered the data gathering process from the middle-ware (sensors) to create timely profiles for each patient representing their performed daily activities.

• Analysed old people behaviours in time and organised non-structured data into a relational PostgreSQL database.

• Modeled normal behaviour profiles of patients to provide the caregivers with a condensed representation of their daily routines.

• Designed a continuous machine learning algorithm in PyTorch to report anomalous behaviours in newly collected sequences of patient activity data.

• Developed periodical code snippets in python to update the patient predictive models saved on Google Drive to reflect for possible seasonal changes in the patients’ routines.

• Guided team members at programming visual plots to assist caregivers at devising prompt intervention strategies for patients with reoccurring anomalous behaviours.