LogSense

LogSense watches incoming log data and flags entries that fall outside a normal statistical range. It uses a Z-score based approach: each value gets compared against the recent mean and standard deviation, and anything far enough from the norm gets flagged as a potential anomaly.

The backend is built with FastAPI and exposes the anomaly detection logic as an API. A Next.js dashboard sits on top of it, so the flagged anomalies actually show up somewhere readable.

The whole thing runs in a Docker container on AWS EC2, with GitHub Actions handling build and deploy on every push, the same CI/CD discipline used across the rest of this portfolio's projects.