Amazon SageMaker Reviews

What I like best about Amazon SageMaker is its ability to manage the entire machine learning lifecycle in one integrated platform. It simplifies model building, training, and deployment while offering scalability and powerful tools like SageMaker Studio and automated model tuning.

What I dislike about Amazon SageMaker is that its pricing can be complex and quickly become expensive, especially for long-running training jobs or large-scale deployments. Additionally, the learning curve can be steep for new users unfamiliar with AWS services and configurations.

Amazon SageMaker solves key problems in data science and machine learning, such as managing infrastructure, automating model training and tuning, and streamlining deployment. This benefits me by accelerating the ML workflow, reducing time spent on setup and DevOps, and enabling faster experimentation and production-ready model delivery.

It is highly scalable, very compute-powerful, very well integrated with most vendors' data warehouses and data lakes, and can be accessed in the browser.

I can hardly make an estimate of the price calculation. Even though there is some tool called AWS pricing calculator, the list of available configurations doesn't show the number of configurations you can select while setting up the tool Studio and Notebook instances.

I use AWS SageMaker daily for data science projects, where Studio and Notebook instances are mainly used as a prime development environment. Now, what makes this tool ideal, due to its being in the cloud, is that you can work around large volumes of data with the ability to scale and have more resources as needed with a simple click.

Provision of built in Algorithms and framework. Lot of the times, it's the data that causes the issues with the predictions. When we got the data right the predictions based on the built-in Algorithms did a great job in linear, logistic, classification techniques. Collaborative with other data scientists. It is easy to integrate with other related systems like Salesforce when we have our data in S3 buckets and the customer support is very responsive.

Though we are getting compute for a reasonable costs, the onus of responsibility to run the large model lies witth the users. When they run larger models just to test it is attracting some additional costs. Though Sagemaker is easy to use, the cost management responsibility lies with the users.

Data

I like the endpoint creation which can inference our model through lambda function. Along with Sagemaker I used API gateway as well as to use the model in local environment.

When I used sagemaker for object detection using py torch then it was not taking the image and I used to resize the image and then pass it to model. But in local it was working without even resizing the image

It can build the model and can be infered after endpoint creation and after that further processing can be done via lambda function and API gateway.

In sagemaker we can have some of basic models and we can create complex AI models and train and test without hassle.

Sagemaker is not having extensive data for images.

Forecasting of future emission levels

SageMaker makes it very easy to train and deploy models. The managed infrastructure allows us to focus on business logic without needing to deal with things like cluster management, autoscaling, etc.

Sometimes things can be a little overcomplicated to use. For example, the batch transform functionality requires us to write separate inference containers in addition to real-time inference containers.

AWS SageMaker helps us build and deploy large-scale recommenders in both offline batch and online ecosystems.

I am exclusively using Amazon SageMaker for both professional and personal usage. The variety of application make Handy while work upon machine learning task. The training and canvas features i've been using for quite some time there application make my ML task faster and productive

Amazon SageMaker is great platform for Ml task all features and applications are really easy to use. The feature I need attention is the the offered free trial which are not sufficient and Amazon should provide gpu access also. Apart from this it's a great ml online platform

I am a data scientist my most of the work are based of ML Modal research and development.

Amazon SageMaker is widely present all over the world with remarkable technical support. Excellent knowledge and details of products and best tracking of shipments.

Restrictions on destinations for certain products for reasons I do not know. But these warnings are informative to avoid losing money.

Amazon SageMaker allows me to discuss in real-time and have advice on products and flexible payment methods with Gift Cards and bank transfers. Speed, efficiency, and reliability of transactions.

Sagemaker was very easy to use and the interface was user friendly

There was no access for the user to check her daily stats. The access was restricted.

Traning the

Amazon Sagemaker is a managed service for data science, We like SageMaker because it supports a lot of in-built algorithms, we especially use liner learner algorithms for solving predictive maintenance use cases

We noticed, as a startup, when we are beginners. If we failed to stop unused sage make endpoint.That will end up in huge bills. Maybe looking for a better solution to fix this issue.

we especially use liner learner algorithms for solving predictive maintenance use cases & we are also using Sagemaker Neo for compiling Sagemaker trined Ml modles.