Designing and Building Complex Machine Learning Projects
summary
Practical recipe for production ML: use SageMaker to spin up training and inference with a few SDK lines, run automated hyperparameter tuning on transient GPU instances, and detect bias with SageMaker Clarify (class imbalance score near 1 flags problems). Combine Step Functions and SageMaker to auto-retrain and promote models, run 80/20 A/B testing on endpoints, or deploy small TensorFlow or scikit-learn models inside Lambda for near-zero cost. Emphasize logs/events, CI/CD pipelines, and minimizing time to deploy.
AI-written summary · about
chapters (5)
- 0:00Start
- 6:25SERVERLESS
- 11:20MACHINE LEARNING ENGINEERING
- 15:13LOGS AND EVENTS
- 32:43PRINCIPLE OF LEAST PRIVILEGE
related talks
33mMachine Learning Engineering Done Right
34mLat Designing and Building Complex Machine Learning....
37mStreamlining Machine Learning Operations
33mBuild, train, and deploy ML models using Amazon SageMaker
37mStreamlining Machine Learning Operations
44mPragmatic Machine Learning in the Cloud
31mBuild, train, and deploy ML models using Amazon SageMaker
34mDeploying a Machine Learning System That Only Runs Locally