G2

Atharva P.

Cloud BI Engineer

Enterprise (> 1000 emp.)

7/29/2026

"End-to-End ML Platform That Streamlines the Full Lifecycle"

4.5/5

What do you like best about Amazon SageMaker?

Amazon SageMaker AI is an end-to-end machine learning platform that brings data preparation, feature engineering, model training, hyperparameter optimisation, deployment, monitoring, and MLOps together in a single managed service. I especially like SageMaker Studio, JumpStart, Pipelines, Feature Store, Training Jobs, Real-Time Endpoints, Batch Transform, Model Registry, and Clarify, as they help streamline the entire ML lifecycle from experimentation through to production.

SageMaker Studio offers a unified workspace for data scientists, and the managed infrastructure significantly reduces operational overhead. Performance is also excellent, thanks to distributed training, managed GPU instances, automatic scaling, and built-in optimisation features that make it easier to run and iterate on workloads efficiently. Integration with Amazon Bedrock further supports hybrid architectures, letting teams combine traditional ML models with foundation models for generative AI applications.

Pricing can be substantial for GPU-intensive workloads, but in my experience the reduced infrastructure management burden and faster path to deployment often translate into strong ROI for enterprise ML teams. Review collected by and hosted on G2.com.

What do you dislike about Amazon SageMaker?

The service has a steep learning curve because it provides a wide range of capabilities across the ML lifecycle. Costs can also rise quickly if notebook instances, endpoints, or GPU resources aren’t monitored and managed carefully. For new users, Studio may feel overwhelming as well, given the sheer number of services and options available. Review collected by and hosted on G2.com.

What problems is Amazon SageMaker solving and how is that benefiting you?

Amazon SageMaker AI solved the challenge of building, training, deploying, and monitoring machine learning models at scale without managing ML infrastructure.

Example: A demand forecasting solution used SageMaker Feature Store to manage training features, SageMaker Pipelines to automate model training, Hyperparameter Tuning Jobs to optimise accuracy, Model Registry for version control, and Real-Time Endpoints for online predictions. Model Monitor continuously detected data drift, enabling the team to maintain prediction quality while reducing manual operational effort. Review collected by and hosted on G2.com.