top of page




How I Deployed My NBA Prediction App Using Hugging Face Spaces and Streamlit
In this final chapter of my NBA Insights project, I walk through the complete deployment process—turning a local machine learning app into a live web application on Hugging Face Spaces. Learn how I handled Docker vs. Streamlit SDK choices, securely managed secrets like Google API credentials, and worked around model file upload issues using Hugging Face Datasets. This post is packed with real deployment lessons, code snippets, and a look ahead at what’s coming next for the NB
Aykut Onat
Jul 272 min read
Â
Â
Â


Building NBA Insights AI — Weeks 3 to 5 Recap
In this three-week sprint, I took NBA Insights from a basic prediction script to a powerful AI system that understands momentum, rest, and game context. From experimenting with Logistic Regression vs. XGBoost to engineering features like win streaks and opponent win rates, this post documents how I trained, evaluated, and deployed an NBA prediction model that hit 70% accuracy on unseen data. Whether you're into machine learning, sports analytics, or building AI tools with rea
Aykut Onat
Jul 62 min read
Â
Â
Â


From Image to SEO: Building a Smarter Way to Write Product Listings
Tired of writing SEO product listings by hand? I built an AI-powered tool that turns product images into Amazon-style titles and bullet points — in seconds. In this post, I share how it works, why it matters, and what’s next for smarter catalog content.
Aykut Onat
May 222 min read
Â
Â
Â


We Used to Stage the Couch. Now We Just Describe the Room.
If you’ve ever helped build lifestyle imagery for a product catalog, you know the pain: Shipping delays, lighting revisions, styling...
Aykut Onat
May 123 min read
Â
Â
Â
bottom of page