I build machine learning systems, data pipelines, and full-stack applications that turn raw data into working products. Based outside Houston, shipping worldwide.
I'm a Python developer and data scientist based in Huffman, Texas. I work at the intersection of software engineering and machine learning — building everything from predictive models and NLP pipelines to the Django and React applications that put them in front of real users.
My toolkit runs deep on the data side: Pandas, Scikit-learn, SQL, and the modeling workflows that go with them — classification, regression, clustering, computer vision, and natural language processing. On the engineering side, I ship production web apps with Django, Flask, and React.
I'm currently available for freelance development and data consulting. If you have data and a problem, I can build the thing that solves it.
End-to-end churn prediction for a telecom operator: EDA surfaced that monthly subscribers and fiber-optic users churn hardest (especially in the first four months), then a model was built to flag at-risk customers for targeted retention offers.
Deep learning model for the Good Seed supermarket chain that estimates a customer's age from a photo to support alcohol-law compliance. Fine-tuned a pre-trained ResNet50 on GPU, testing dropout and learning-rate configurations.
Automated sentiment classifier for the Film Junky Union, trained on labeled IMDb reviews. Compared NLTK vs. spaCy lemmatization, TF-IDF features, logistic regression, gradient boosting, and DistilBERT — beating the required F1 threshold.
Hourly demand forecasting for Sweet Lift Taxi's airport fleet. Resampled and decomposed historical order data, engineered time features, and benchmarked models — LightGBM won, letting dispatch staff drivers ahead of peak hours.
Profit-optimization study for OilyGiant: regression models predicted reserve volumes across three regions, then bootstrapping with 1,000 samples quantified expected profit, 95% confidence intervals, and downside risk before recommending a region.
Interactive web app for exploring US car-sales listings — histograms, scatter plots, and filter controls built with Streamlit and Plotly Express, version-controlled on GitHub and deployed live to the cloud on Render.
Need a machine learning model, a data pipeline, or a full-stack application built? I'm available for freelance development and data consulting. Send a transmission.