CourseAI / ML Engineer: The Complete Course · Module 2: Software and Systems Engineering · part 15 of 93
Part 15 · Module 2: Software and Systems Engineering

Lesson 2: The systems around the code

29 min read·
Module 2 · Lesson 2 · About 55 minutes · Needs: lesson 1's idea of a package and tests (its files are not required), and the Kestrel data (course-guide/06).

The situation

Week 2 continues. Rhys owns the 06:00 batch, which must score every parcel due that day within 20 minutes (an illustrative target). Adaeze asks: "When we ship a model file, how do we know loading it is safe, and that next month's code can still read it?"

The running example is the 06:00 batch, and inside it KP0104204 (simulated): scored over the network, written in five formats, and saved as a model file.

WHERE THIS FITS The Kestrel system this lesson SHARED SERVICES Depot scans parcels, routes Ingestion nightly batch Validation Pandera checks Data lake Parquet, DuckDB Experiments MLflow runs Training LightGBM, Optuna Online store Feast, Redis Features point in time Gates vs the champion Registry staged versions Serving API FastAPI, ONNX Consumers vans, customers Retraining trigger, budget Alerting on call Monitoring drift, Evidently Late labels a day later Orchestrator Prefect flows Versions DVC, Git CI/CD GitHub Actions Containers Docker images
This lesson adds the lake's file format, version control, CI and the container image around Kestrel's code (blue); the rest is grey for now.

The rest of this course is yours to keep

This course is bought on its own, once, and stays readable afterwards, including the parts added to it later.

Buy this course · ₹399See what is inside

Or open everything with a plan