Data Science
Resume:
Joshua Felce
***********@*****.***
https://www.linkedin.com/in/joshuafelce/
https://github.com/JF400
Skills
• Python (OOP, Pandas, NumPy, Matplotlib,
Seaborn)
• SQL (MySQL, PostgreSQL, SQL Server)
• Apache Spark (PySpark, Spark DataFrames,
MLlib, Spark SQL, RDD, Streaming)
• Git/GitHub/GitHub Desktop
• Deep Learning (Keras, TensorFlow)
• Machine Learning (Scikit-Learn)
• Statistics (ANOVA, T-tests, PCA, Bayesian Inference)
• Excel
• Calculus/Linear Algebra (Matrices, Gradient Descent)
• English (native), Spanish (native)
Education
MASTER OF SCIENCE IN DATA SCIENCE – University of Denver, CO, USA – GPA: 4.0/4.0 2023 – 2025 Courses: Database Organization & Management, Parallel & Distributed Computing, Python Software Development, Deep Learning, Machine Learning, Probability and Statistics 1 & 2, Algorithms for Data Science, Data Visualization BACHELOR OF SCIENCE IN MATHEMATICS AND PHYSICS – University of Calgary, AB, Canada – GPA: 3.5/4.0 2018 – 2023 M.S., Data Science Projects
RETINAL SCAN MULTI DISEASE CLASSIFICATION April 2025 – June 2025
● The dataset consisted of 3,544 retinal scan images, each annotated for the presence of 7 eye diseases. A retinal scan may be associated with multiple diseases simultaneously.
● The task was to build a multi-label classifier that identifies the specific eye diseases present in a scan. Two different approaches were employed to achieve this task: The binary relevance method from the scikit-multilearn library and a single convolutional neural network designed for multi-label classification.
● The first technique decomposes the multi-label task into 7 independent binary classification problems—training a separate convolutional neural network (deep learning) and a separate support vector machine or random forest (machine learning) for each disease. F1 threshold maximization was employed for this technique.
● The second technique involved a single convolutional neural network with 7 neurons in the output layer—one for each disease— that were activated by the sigmoid function to produce independent probabilities for each disease.
● The project showcases the use of Python, Deep Learning (Keras, TensorFlow), and Machine Learning (Scikit-Learn). VIDEO GAME DATABASE November 2024 – December 2024
● Designed a video game database model: A Star Schema consisting of a central fact table surrounded by multiple dimension tables.
● The data being modeled consisted of the name of the video game, release year, platform, publisher, genre, and sales numbers in various regions of the world.
● Created 4 stored procedures using MySQL to answer common queries on the dataset. BOOK DATABASE October 2024 – November 2024
● Created a Graphical User Interface (GUI) using Python and SQL where the Python library Tkinter was used to create the GUI and connected to a MySQL database using the Python Connector Interface.
● Users can interact with the GUI to search for books by author, category, or publisher. Work Experience
UNDERGRADUATE STUDENT RESEARCHER – University of Calgary April 2021 – September 2021
● Awarded a grant to conduct and then write a report on a research project done within the Department of Mathematics.
● Used Python to implement two numerical methods to solve a Partial Differential Equation: Centered Finite Difference Method and a Physics Informed Neural Network (built using Keras and TensorFlow). Work Authorization
Authorized to work in the U.S. under OPT (36 months); TN visa available for continued employment (Canadian Citizen).
Contact this candidate