Pedro Filipe Custodio

Hi, I'm an IT Consultant at Findwise. Previously I worked as a Data Scientist at Dynasource and Quantillion. My ambition is to apply data science methodologies in the health care sector.
Before that, I was a PhD fellow researcher at the MOVE Research Institute, partnering with Motekforcelink. I have also conducted research at the LASIGE and University of Lisbon.

I do research, I did some teaching, I am learning more computer science, I go to the gym and I go to bootcamp.

Work


At Findwise, I work as an IT Consultant and I have been involved in multiple projects that require different set of skills.

IT Consultant (April 2019 - now)

In one of the projects for Autoliv, I work as a frontend developer using both Angular and Vue. I also been in close collaboration with backend developers which helped me in learning Java. At CEVT I worked mostly with Python developing machine learning models, deploying APIs and using Azure Functions to make them available. And the final project is with SKF where I use AWS Glue to process data and create tables to use on PowerBI.

At Dynasource, I worked in an agile framework, where I belonged to the data scientist team. We collected and analysed users’ data inside a web platform for IT professionals and service providers claim skills.

Data Scientist (2018 - March 2019)

We were developing Natural Language Processing (NLP) models and Product Recommendation systems in Python. Other tools commonly used in the team were Spark, Microsoft Azure and Elasticsearch. We also developed a small web app using Flask. The web app was able to send requests to our database using authentication and present in a table the data to edit it.

At Quantillion, I worked as a Junior Data Scientist Consultant. I analysed data to help organizations solve their operational and strategic challenges.

Junior Data Scientist (2017 - 2018)

I was involved in projects using JavaSript visualizations and graphical databases technologies like Neo4j. I worked with different kinds of data, e.g., financial datasets. During this time, I was introduced to a variety of back- and front-end tools like Amazon Web Services (AWS), Mongo dB, Neo4j, Python, R, Angular and JavaScript.

Research

EEG recording
during walking

In EEG recording during walking: a literature review, I review 43 articles that used EEG methods to study the brain during movement of lower limbs. I summarized the preprocessing methods to clean EEG data of artefacts (mostly muscular and movement related) and the analyses methods used to extract meaningful measures of the brain. I concluded that ERSPs and ERDs on the time-frequency and ERPs on the time analysis were the most used.

EEG recording during walking: a literature review

Understanding the importance of cognitive processes in locomotion is one of the challenges of cognitive neuroscience, mainly due to methodological limitations of neuroimaging techniques (R. Holtzer, et al., 2014; S. Makeig, et al., 2009). Studying the brain during whole body interactions in a non-controlled natural environment is not always possible. Recently, several studies have been using EEG to study brain activation patterns during locomotion. However, some discussion around the propagation of muscular artifacts on the EEG signal suggest that a more careful and thoroughly look should be taken when analyzing those results (M. Wieser, et al., 2010). The aim of this review is to provide more information on the EEG methodologies state of the art used to study gait control in humans. Thus, we reviewed existing articles that apply EEG to study locomotion in regards to the methodologies, hypotheses (whether it was focused on validation of methodologies or population groups’ research, etc.), applications and future potential of this neuroimaging technique.

Stability outcomes in Parkinson's Disease

In Stability outcomes in Parkinson's disease patients we compare the gait data of 12 PD patients and 12 control participants during walking tasks under different visual and auditory conditions.

Stability outcomes in Parkinson's Disease

In Stability outcomes in Parkinson's disease patients we compare the gait data of 12 PD patients and 12 control participants during walking tasks under different visual and auditory conditions. We concluded that stability was different in PD patients between visual cueing and no cueing and between auditory cueing and no cueing. Control patients had better stability results compared with PD group in all conditions.

Effects of Visual
Perturbations in PD patient

In Effects of Visual Perturbations in PD patients we analyze the gait data of 12 PD patients in conditions with visual manipulation of the screen. The objective of the project was to understand the effects of visual manipulation on PD patients' gate and how the cortex activation were altered.

Previous research

To read or not to read:
a neurophysiological study

In To read or not to read: a neurophysiological study, we report a study case about a patient with pure alexia, in whom event-related potentials (ERPs) revealed a delay in the P100 component and an absent N170 compared to controls. Thi suggests that in this patient case the alexia might result from deficits in visual perception which leads to an inability to activate the VWFA.

To read or not to read: a neurophysiological study

Pure alexia (PA) has been associated with visual deficits or a failure to activate the visual word form area (VWFA). We report a patient with pure alexia due to posterior cortical atrophy, in whom event-related potentials revealed a delay in the P100 component and an absent N170 compared with controls. Furthermore, there was a tendency for a larger delay in P100 latencies associated with incorrectly read words. This suggests that some cases of PA might result from deficits in visual perception, signaled by the P100 early potential which could lead to an inability to consistently activate the VWFA, marked by the absent N170.

Education

PhD Research Fellowship (August 2013 to 2016)

Master in Biomedical Engineering (September 2008 to October 2010)

BSc. in Biomedical Engineering (September 2005 to July 2008)

Other
Email

Feel free to contact me via email at Mail or using the following form: