Full-stack · Service prototype
AeroOps Java
A Spring Boot flight-delay prediction service with Java XGBoost inference, Redis caching, and PostgreSQL audit logs.
JavaSpring BootRedisPostgreSQL
Explore repository 
The project
A Java serving companion to AeroOps ML. A native tree evaluator reads the XGBoost JSON model, including categorical splits and missing-value branches, without a Python sidecar. The service adds request validation, Redis cache-aside, asynchronous PostgreSQL audit writes, and OpenTelemetry instrumentation. Repository tests compare predictions with Python-generated fixtures. It is a local service prototype with no verified cloud deployment, authentication, or rate limiting.
Repository activity ·
Inside the build
- 01Pure Java XGBoost tree evaluation with numerical and categorical splits
- 02Redis caching with an inference fallback when the cache is unavailable
- 03Asynchronous PostgreSQL audit writes and Flyway migrations
- 04OpenTelemetry instrumentation and Python-to-Java prediction parity tests