Anyscale Managed Ray Platform
AnyscaleReviews from AWS customer
0 AWS reviews
-
5 star0
-
4 star0
-
3 star0
-
2 star0
-
1 star0
External reviews
5 reviews
from
External reviews are not included in the AWS star rating for the product.
Great tool for scaling AI workloads
What do you like best about the product?
What impresses me most is how it handles the heavy lifting for Ray. I can develop my AI application code right on my laptop and then deploy it to a large cluster without having to rewrite anything or wrestle with complex infrastructure setups. This effectively bridges the gap between code that only "works on my machine" and a real production environment, which is particularly useful when scaling LLM workloads and managing distributed training. In the end, it saves me a considerable amount of time on DevOps tasks.
What do you dislike about the product?
The pricing structure can feel somewhat unclear, making it difficult at times to anticipate your final monthly bill. This is especially noticeable when compared to the more straightforward cost management you get with handling raw EC2 instances on your own.
What problems is the product solving and how is that benefiting you?
I use Anyscale mainly to overcome the infrastructure challenges of scaling Python machine learning code from my local laptop to a large distributed cluster. My team operates a substantial Retrieval-Augmented Generation (RAG) pipeline, which includes OCR processing and embedding generation for millions of PDF files. Previously, running this workload on a single large EC2 instance would take weeks, and managing AWS Batch jobs involved a lot of boilerplate and ongoing DevOps work. With Anyscale, we were able to wrap our existing Python functions with Ray decorators, enabling the platform to automatically spin up a cluster of over 50 spot instances, process 2TB of data in less than four hours, and then scale back down to zero. This approach has reduced our compute costs by about 60% by taking advantage of spot instances without the need for manual fault-tolerance solutions, and it has allowed my data scientists to independently run large-scale experiments without waiting for DevOps to provision resources.
AI/ML
What do you like best about the product?
Anyscale makes it easy to scale AI/ML workloads without worrying about infrastructure complexity
What do you dislike about the product?
Documentation could be more beginner-friendly with clearer end-to-end examples
What problems is the product solving and how is that benefiting you?
solves the challenge of scaling machine learning and AI workloads without requiring deep expertise in distributed systems. eg remove complexity
Scalable and reliable platform for AI workloads
What do you like best about the product?
Anyscale simplifies the process of moving AI and ML workloads from development to production. Since it is built on Ray, it enables scalability without requiring major code changes.
What do you dislike about the product?
The platform has a noticeable learning curve, particularly for teams unfamiliar with Ray concepts. Pricing is not always transparent, which makes cost planning more challenging.
What problems is the product solving and how is that benefiting you?
Anyscale addresses the challenge of running distributed ML and GenAI workloads efficiently.
Good
What do you like best about the product?
The Anyscale platform was essential as it fully managed, production-ready version of Ray, offering a simplified and integrated developer experience and it made easy to build and it has good scalability.
What do you dislike about the product?
About the disadvantage is during building there is little trouble when debugging trouble.
What problems is the product solving and how is that benefiting you?
I solved the scalability and robustness problems as it was easier to solve.
Infrastructure for AI
What do you like best about the product?
It's provide infrastructure for AI and deep learning.
What do you dislike about the product?
I haven't found anything wrong with the product.
What problems is the product solving and how is that benefiting you?
We were struggling with the risk analysis for the wind turbines components but with the help of Anyscale ray technology we Easley crack it with high true rate.
showing 1 - 5