MLops: The Key to Pushing AI into the Mainstream

We’re excited to bring Transform 2022 back to life on July 19th and virtually July 20-28. Join AI and data leaders for sensible conversations and exciting networking opportunities. Register today!


One of the main obstacles preventing an enterprise from implementing Artificial Intelligence (AI) is the transition from development and training to the production environment. To reap real benefits from technology, this must be done at the speed and scale of today’s business environment, which few organizations are capable of doing.

This has led to growing interest in merging AI with devops. The forward-learning enterprise is typically trying to blend machine learning (ML) with the traditional Devops model, creating an MLops process that streamlines and automates the way intelligent applications are developed and deployed and then continues to add value. Updated on a regular basis. Its operation over time.

Problem solver

According to data scientist Ayman Hutchman, MLops help enterprises deal with a number of significant issues while effectively creating and managing intelligent applications. For one thing, the data sets used in the training phase are extremely large and are constantly expanding and changing. This requires constant monitoring, experimentation, adjustment and retraining of AI models, all of which are time consuming and costly under traditional, manually driven development and production models.

To implement MLops effectively, the enterprise will need to develop a number of key capabilities, such as full lifecycle tracking, metadata optimized for model training, hyperparameter logging and solid AI infrastructure that includes not only server, storage and networking solutions but also software. . Rapid repetition of new machine learning models. And all of this has to be designed around two main forms of MLOPS: forecasting, which seeks to chart future results based on past data and prescriptions, which seeks to make recommendations before decisions are made.

Mastering this discipline is the only sensible way for AI to move from Fortune 500 Enterprise to the rest of the world, says Greenfield Partners’ Shay Greenfeld and Itai Inbar. In fact, 90% of ML projects fail under the current development and deployment framework, which is not suitable for most organizations. MLops provide a dramatically more efficient development pipeline that not only reduces the overall cost of the process but can quickly turn failure into success. The end result is that barriers to AI implementation go to a level that is comfortable for most enterprises, leading to widespread integration and ultimate integration into mainstream data operations.

Early success

Sibanjan Das, a business analytics and data science consultant, says MLops is still an emerging field, so he may be tempted to write it as another tech buzzword. But its track record so far has been very good, even if it is properly designed and aimed at the right goal: to maximize model performance and improve ROI. This requires careful coordination between the various components that create the MLops environment, such as the CI / CD pipeline, as well as model servicing, version control, and data monitoring. And don’t forget to create strong security and governance mechanisms to reduce the risk and potential for the ML model’s activities to be compromised.

Although MLops are also designed for automation and autonomy, do not overlook the human element as the main driver of successful results. A recent report by Dataiku noted that over the past year, companies have realized that they cannot scale AI without creating different teams that can implement and leverage the technology. MLops should be an important component of this strategy as it supports diversification in the development, deployment and management of AI projects. And just considering Gartner’s MLops framework, a broad set of skills will be needed to ensure that results deliver high value to the enterprise business model.

Even the most advanced technology is less valuable if it cannot successfully transition from a laboratory to the real world. AI is now at a stage where it must begin to make a valuable contribution to humanity or it will become the digital equivalent of AdSell: flashy and full of gadgets but with less practical value.

Of course, MLops can’t guarantee success, but it can reduce the cost of experimentation and failure, while at the same time putting it in the hands of more people who can figure out for themselves how to use it.

Venturebeat’s mission Transformative Enterprise is about to become a digital town square for technology decision makers to gain knowledge about technology and transactions. Learn more about membership.

Similar Posts

Leave a Reply

Your email address will not be published.