I built a real‑time ML feature pipeline that computes 15+ features with sub‑100ms latency.
Stack: Go ingestion → Kafka → Python feature processor → Redis cache + TimescaleDB store → Feature API.
It includes feature versioning, A/B testing, drift detection, DLQ, Prometheus, and Grafana dashboards.
Run locally with Docker Compose and verify via included test scripts.
https://github.com/judeszn/Real-time-ML-Feature-Pipeliine-
Top comments (0)