EMA Releases New Top 3 Decision Guide for MLOps to Assist Data Scientists, Data Engineers, Software Developers, IT Operators, DevOps Engineers, and Security Professionals

By STS News Desk On Jan 18, 2022

EMA-Releases-New-Top-3-Decision-Guide-for-MLOps-to-Assist-Data-Scientists_-Data-Engineers_-Software-Developers_-IT-Operators_-DevOps-Engineers_-and-Security-Professionals

New research from EMA highlights platforms that enable organizations to accelerate the creation, operationalization, and continuous enhancement of machine learning-driven applications

Enterprise Management Associates (EMA), a leading IT and data management research and consulting firm, announced it has released a new report titled “EMA Top 3 Decision Guide for MLOps in 2022,” authored by Torsten Volk, managing research director of cloud-native, DevOps, machine learning, and AI. EMA selects its Top 3 products based on the analysis of priorities and real-world challenges experienced by data scientists, data engineers, software developers, IT operators, DevOps engineers, and security professionals.

The EMA Top 3 report enables enterprises to benefit from the experiences, good and bad, of their peers when infusing their application portfolio with machine learning capabilities that work across data center, public cloud, and edge locations.

Enabling the entire organization to benefit from machine learning capabilities across its overall application portfolio is a critical pain point for this group of professionals in 2022. This focus is crucial because data scientists struggle to efficiently create impactful machine learning models in a world of inconsistent data access and data science tooling, while data engineers spend too much time building and maintaining the underlying infrastructure platforms required for the optimal performance and reliability of data pipelines and machine learning models. Developers and DevOps teams often lack the ability to effectively incorporate machine learning models into their applications.

“The EMA Top 3 report enables enterprises to benefit from the experiences, good and bad, of their peers when infusing their application portfolio with machine learning capabilities that work across data center, public cloud, and edge locations,” says Volk. “The report provides decision-makers and influencers with concise and up-to-date market data and a shortlist of products that address the pain points and priorities reflected in this data.”

Some of the key findings are:

Cost is the #1 infrastructure-related challenge for machine learning for 35% of organizations. Scalability is the biggest headache for 28%. Software cost makes up 38% of cost-related challenges, with the implementation of new technologies and new systems both coming in at approximately 16%.
Data-related issues and model building make up 55% of scalability challenges.
Clustering and classification models are most affected by infrastructure-related challenges. Pytoch and TensorFlow are the two critical deep learning frameworks experiencing issues within an infrastructure context.
Feature-related topics take up ranks #2, #5, #13, #19, #27, and #28 in our list of the top 30 data science challenges in 2022.
Performance, scalability, and cost take up the first three spots related to the critical machine learning challenges within a Kubernetes infrastructure context.

A detailed analysis of EMA’s research is available in the “EMA Top 3 Enterprise Decision Guide for MLOps” report.

Write in to psen@itechseries.com to learn more about our exclusive editorial packages and programs.