Felipe Andrade Rubio

Junior Data Scientist

Ama Earth Group•  April 2024 - July 2024•  Remote

Conducted Exploratory Data Analysis (EDA) using Python, NumPy, Matplotlib, and pandas on 5+ databases. Integrated Google Analytics on the company’s website to track user and website activity using React and HTML. Extracted, Transformed, and Loaded (ETL) geospacial data from Google Earth Engine using JavaScript. Developed a geospatial React multi-threaded web app for classifying mangrove areas by health with a decision tree, using Google Earth Engine, Google Cloud, Greppo, REST API, Scikit-learn, and Python with 95% accuracy, deployed using Docker.

Research Assistant

The Pennsylvania State University•  May 2023 - December 2023•  Pennsylvania, US

Researched with Ph.D. student Yitzchak Shmalo on Machine Learning and Deep Neural Networks, using Python, PyTorch, and NumPy in Google Colab. Programmed and tested more efficient Deep Learning models using Random Matrix Theory to reduce the number of parameters and increase performance for ResNet, VGG, DNN, and CNN with accuracies of 95+%. Applied the splitting method of Random Matrix Theory to the regression problem and measuring Deep Neural Networks performance; achieved an increase in accuracy of 5+% on average.

Software Developer

Penn State Advanced Vehicle Team•  August 2022 - September 2023•  Pennsylvania, US

Collaborated with a team of 60+ students to develop a L4 Autonomous Car in Data Acquisition and Simulation for the SAE AutoDrive Challenge II, programming in Python, C, YOLOv7, Webots, and SUMO. Extracted, Transformed, and Loaded (ETL) segmentation data of lane markings from 100,000 images of the BDD100k database. Developed YOLOv7 lane detection model for in-track testing, using instance segmentation for lane type recognition on the road. Constructed the M-City Road network and 3+ test tracks in Webots and SUMO for testing, prototype design, and use in CARLA. Implemented Prototype Car with C controller, LIDAR, radar, camera, speedometer, GPS, and sensors in Webots. Placed third in Year Two of SAE AutoDrive Challenge II out of 10 universities.

Software Development Intern

InnScience•  June 2021 - August 2021•  Remote

Created a website prototype with HTML, CSS, and Python tested in Flask. Integrated 3+ web pages into InnScience’s website and made a general template for HTML files using Jinja. Proposed and discussed new ideas for implementation into InnScience’s website.