machine-learni

End-to-end Machine Learning project for customer credit risk prediction using PostgreSQL, XGBoost, FastAPI, and GitHub.

Autonomous, AI-driven Site Reliability Engineering (SRE) platform. Proactively monitors GitHub repositories, deeply analyzes code diffs using a custom fine-tuned LLaMA model, and predicts runtime crashes before they hit production.

RNN based prediction of customer churn at a supermarket

Universal checkpointing for C# machine learning. Persist full training states (Weights + Optimizer + Hyperparameters) to File System, PostgreSQL, or Cloud Blobs. Manage thousands of models by GUID with zero framework lock-in—built for custom training loops first.