AI
Data-driven game predictions
ML model for game predictions with a Next.js dashboard and Recharts for accuracy and trend analysis. Python backend for training and inference.
The client wanted to offer NBA game predictions backed by data, with a way to monitor model performance and explain results to stakeholders. They had historical data but no pipeline or interface to turn it into a product.
We built a Python pipeline for feature engineering and model training, an inference API for game-day predictions, and a Next.js dashboard with Recharts to track accuracy, compare to baselines, and visualize trends over time.
“We had the data; Saints turned it into something we could ship and improve on.”