AWS · MLA-C02 · BETA
Training jobs, pipelines, and retraining
MLA-C02 · Lesson 17 of 23 · 2:14:24
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Lesson 17 of the free MLA-C02 beta series (AWS Certified Machine Learning Engineer – Associate, domain 3, deployment and orchestration). In this lesson: training jobs, transform jobs, and Step Functions, SageMaker Pipelines with quality gates and caching, retraining triggers, and Model Registry, MLflow, lineage, and versions of fine-tuned models. With three quick checks and a recap, and AnyCompany Mobility, a car sharing service, as the running example.
Preview the topics
- The team's pipeline registers every new model, even when its accuracy is lower than the model in production. The team wants weak models stopped automatically, before registration. What should the ML engineer add?
- New labeled trip data arrives in Amazon S3 at random times, a few times a month. The model should be retrained soon after each delivery, with no one watching. What should start the pipeline?
- An auditor asks which dataset and which training job produced the model behind a production endpoint, six months ago. What should the ML engineer use to answer?