I build projects across analytics, machine learning, and data systems. I'm especially interested in work where real-world data, models, and practical decision-making all come together.
A mix of analytics, data systems, machine learning, and computer vision projects built around making messy information more useful.
A Python machine learning pipeline that audits decision quality, engineers structured features, trains a classifier, and validates improvements through shadow evaluation. Built on Pokemon battle data as a real-world audit exercise.
A searchable inventory system for tracking graded trading cards, sale records, and collection data using SQL and Python.
Built a computer vision classifier in PyTorch that uses a convolutional neural network to classify CIFAR-10 images into 10 categories. The project includes model training, test accuracy, a confusion matrix, and sample prediction outputs.
An interactive dashboard built in Power BI to visualize web traffic patterns, session behavior, and engagement metrics. Includes data cleaning in Python and a structured layout designed for non-technical stakeholders.
What I'm working on right now — including a sports prediction model, an MLB data pipeline, and ongoing skill-building work.
Building across analytics, machine learning, and data systems. More on GitHub.
I recently finished my M.S. in Information Science and Technology at UW-Milwaukee. The program focused on data science, analytics, and information security, with a lot of work around large datasets, structured systems, and solving real problems with data.
I also bring client-facing experience from photography, videography, hospitality, and sales roles, so I care just as much about communication and follow-through as I do about the technical side.
Most of my experience comes from hands-on projects where I've worked with SQL, Python, Power BI, and databases to clean data, build reports, and organize information in a way that supports better decisions. That has included designing relational databases, building dashboards, connecting APIs, and writing the ETL and validation work needed to keep data reliable.
More recently, I've been building deeper into machine learning, computer vision, and predictive modeling through projects that turn messy raw inputs into structured features, train models, and evaluate where those models actually help. What I like most is combining a strong data foundation with the modeling side instead of treating them like separate tracks.
M.S. in Information Science & Technology from UW-Milwaukee.
SQL, Python, LightGBM, PyTorch, scikit-learn, Power BI, MySQL, R.
Built a walk-forward validated MLB prediction model, an XGBoost NBA props system, and a deployed CNN image classifier.
If you want to connect, ask about a project, or talk about a potential opportunity — feel free to reach out directly.
Or reach me directly at obrad.matthew@gmail.com