CourseAI / ML Engineer: The Complete Course · Module 13: Specialisation · part 78 of 93
Part 78 · Module 13: Specialisation

Lesson 1: Nine specialist tracks, and a taste of each

29 min read·
Module 13 · Lesson 1 · About 45 minutes · Needs: Module 3 (LightGBM, PR-AUC), Module 4 lesson 3 (CNNs, the message classifier) and the Kestrel data (course-guide/06). Lesson 2 then takes one track to depth.

The situation

Week 13 at Kestrel. Adaeze has budget for one specialist hire next year and asks you which kind: "Language? Images? Ranking? Or more of what we do now?" Hiring is organised around tracks, so the answer is one.

One parcel runs through the lesson: KP0099843 (simulated), collected at Dartford on the 31 August bank holiday, route DAR-R06, promised 2 September. It was late, and its customer wrote message M0980: "you said on monday for kp0099843. it is now late?" Each track sees it differently.

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
A track changes what is trained and how it is judged (blue); the data path, serving and monitoring around it carry over (grey).

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