Bobby Lindsey

Join us in working through the ModelOps lifecycle, starting with experiment, train, register, approve, to deploy, anchored by the Amazon SageMaker catalog system of record.

As AI development proliferates across enterprise environments, efficiently evaluating LLMs at scale has become a core challenge for data scientists and ML engineers.

Through guided exercises in Amazon SageMaker Studio, you’ll learn how to build agents, evaluate agents with heuristic and LLM-as-judge metrics using MLflow, deploy agents on Amazon Bedrock AgentCore, and implement observability measures using MLflow’s agent tracing.

Amazon SageMaker offers a fully managed MLflow capability, helping you to manage the entire ML lifecycle more efficiently. In this chalk talk, learn how to effortlessly set up and manage MLflow Tracking Servers with just a few steps, streamlining the process and boosting productivity.

Machine learning operations (MLOps) tools help you automate and standardize the ML lifecycle to productionize ML models faster without compromising model performance. Amazon SageMaker provides a breadth of mission-ready MLOps tools to experiment, train, test, deploy, and govern ML models at scale.

MLOps tools help you automate and standardize testing, managing, deploying, and monitoring hundreds to thousands of machine learning models. In this workshop, learn how you can use Amazon SageMaker MLOps capabilities such as Pipelines, Projects, Model Registry, Model Monitor, and CI/CD to deliver new models in production faster while maintaining model performance in production.

Before using a dataset to train a model, you need to explore, analyze, and preprocess it. During this chalk talk, learn how to use Amazon SageMaker to complete these tasks in a Jupyter notebook environment.

Data scientists apply feature transformation on organizations’ raw data to generate aggregated features for their machine learning algorithms. These aggregated features are critical for real-time applications, such as fraud detection, real-time recommendations, and personalization.