The operations half of the ML Engineer role: running models in production reliably. It’s “DevOps for ML,” so it cross-links the DevOps domain for the shared infrastructure (Docker, Kubernetes, CI/CD, Terraform, Prometheus) and focuses here on what’s ML-specific. Part of the ML Engineer track.

  1. MLOps Fundamentals — [Advanced] — the ML lifecycle, MLOps principles, and what ML adds beyond ordinary DevOps (data + models as versioned artifacts, not just code)
  2. Experiment Tracking & Pipelines — [Advanced] — MLflow/DVC/W&B, model registries, feature stores, and data/training pipelines (Airflow, Kafka, Spark)
  3. Monitoring & Edge — [Advanced] — drift detection, model monitoring, and deploying to constrained/edge devices (TFLite, PyTorch Mobile, Jetson)
  4. Serving & Operations — [Advanced] — getting predictions out of a model: serving, versioning, A/B rollout, retraining pipelines, training-serving skew
↑ Back to 02 — ML Engineer