Personal details

Felippe M. - Remote data scientist

Felippe M.

Based in: đŸ‡§đŸ‡· Brazil
Timezone: Brasilia (UTC-3)

About

I'm an Engineer, with a background in Artificial Intelligence, and I've always sought to use technology to solve problems.

In my Master's, my study focused on Artificial Neural Networks, and I've published scientific articles and produced a dissertation on the subject. In the last four years, I've dedicated myself to studying Data Science (Python, ML, SQL, Spark).

I'm responsible and a team worker. During my career, I've taken on several managerial roles, having even led remote teams.

I know how to work with other people and try to treat everyone with respect and courtesy. I'm always open to hearing suggestions, criticisms, and other perspectives.

I have just finished the Data Science Bootcamp with the french school Le Wagon. It was an amazing experience and I was able to further consolidate my skills in Data Science: statistics and probability, SQL, Python (Pandas, Numpy, Sklearn, PyTorch, Tensorflow/Keras), NLP, MLOPs basics (Google Cloud).

I am searching for new challenges and experiences, and I believe I can make valuable contributions.

Work Experience

Tax Inspector and Data Scientist
RECEITA FEDERAL DO BRASIL | Jul 2010 - Present
Python
SQL
Machine learning
Data analysis
Data Science
Apache Spark

2013 - 2016: Head of the Division responsible for the development of the internal framework of the Receita Federal do Brasil's main Business Intelligence's software. The software was mentioned by OECD in the "2022 Comparative Information on OECD and Other Advanced and Emerging Economies" (https://www.oecd-ilibrary.org/taxation/tax-administration-2022_1e797131-en).

2016 - 2019: Head of regional Tax Inspection Division, having managed around 50 people (remote and local) in the period.

Currently, I've been working daily with data analysis, using especially Python and SQL to identify taxpayers with the highest probability of having committed tax fraud or tax evasion.

Sales Manager
PROCTER & GAMBLE | Jun 2005 - Jul 2008
Excel
Data analysis
Data warehouse

Although not directly related to technology, this experience allowed me to develop valuable skills: I was responsible for over US$10 million in annual sales. So, I know what it means to work for a large company and all the responsibilities involved. I also learned a lot about negotiation and customer relations.

At P&G, I stood out for constantly working with customer Data Analysis. I've learned about the company's Data Warehouse and started to develop reports and dashboards for my team.

This knowledge allowed me to better manage my customers' inventory and the possibility of building forecasts (using Excel spreadsheets, at the time) and more efficient sales strategies.

Projects

Compliance Risk Management
2021
Python
SQL
Machine learning
Data analysis
Data Science
Apache Spark
Following recommendations from institutions such as the OECD (Organisation for Economic Co-operation and Development) and CIAT (Inter‑American Center of Tax Administrations), the Receita Federal do Brasil (Federal Revenue Service of Brazil) developed a Compliance Risk Management model. During the project, SQL was used to extract data from the company's Data Lake (Big Data, using Apache Spark), and Python to execute the Data Analysis and build Machine Learning models - different models were developed for each dimension such as tax collection and inspection. For confidentiality reasons, it is not possible to disclose more information about the project.
SISEN
2019
Python
Machine learning
Data analysis
Data Science
SISEN is a federal tax exemption system for the acquisition of new vehicles for taxi drivers and people with special needs. The system has a period of 72 hours to decide (automatically) whether or not the user is entitled to tax exemption. Experience, however, demonstrates that the system's algorithm had weaknesses that made fraud possible, requiring human intervention in these cases. In the project, we used Python to execute the Data Analysis and to build a new Machine Learning model for SISEN (using K-Means), to group users with similar fraud characteristics. For confidentiality reasons, it is not possible to disclose more information about the project.

Education

LE WAGON
DATA SCIENTIST - BOOTCAMP・DATA SCIENCE
Sep 2022 - Nov 2022
PONTÍFICA UNIVERSIDADE CATÓLICA DE MINAS GERAIS
POSTGRADUATE・DATA SCIENCE AND BIG DATA
Nov 2020 - Apr 2021

Certifications & Awards

Microsoft Certified: Azure AI Fundamentals
Microsoft | Sep 2023
Spark with PySpark
UDEMY | Apr 2023