We’re excited to bring back Transform 2022 in person on July 19 and virtually July 20-28. Join AI and data leaders for insightful talks and exciting networking opportunities. Sign up today!
One of the main obstacles that prevents the company from implementing artificial intelligence (AI) is the transition from development and training environments to production environments. To reap real benefits from technology, this must be done at the speed and scale of today’s business environment, something few organizations are capable of doing.
This is why the interest in merging AI with DevOps is gaining steam. Forward-thinking companies are trying to blend machine learning (ML) in particular with the traditional DevOps model, creating an MLops process that streamlines and automates the way intelligent applications are developed, deployed, and then deployed. continuously updated to increase the value of your operations over time.
According to data scientist Aymane Hachcham, MLops helps the enterprise deal with a number of major issues when it comes to building and managing smart applications effectively. For one thing, the data sets used in the training phase are extremely large and are continually expanding and changing. This requires constant monitoring, experimentation, tuning, and retraining of AI models, all of which are time consuming and costly in traditional, manually driven development and production models.
To implement MLops effectively, the company will need to develop a number of core capabilities, such as full lifecycle tracing, optimized metadata for model training, hyperparameter logging, and a robust AI infrastructure that consists of not just software solutions, server, storage and networking, but also software tools capable of rapid iteration of new machine learning models. And all of this will need to be designed around the two main forms of MLops: predictive, which attempts to plot future outcomes based on past data, and prescriptive, which strives to make recommendations before decisions are made.
Mastering this discipline is the only plausible way for AI to trickle down from the Fortune 500 to the rest of the world, say Shay Grinfeld and Itay Inbar of Greenfield Partners. The fact is that over 90% of ML projects fail in current development and deployment frameworks, which is simply not sustainable for the vast majority of organizations. MLops provides a much more efficient development pipeline that not only reduces the overall cost of the process, but can also turn failures into successes at a rapid rate. The end result is that the barriers to AI implementation drop to a comfortable level for the vast majority of companies, leading to widespread distribution and eventual integration into mainstream data operations.
MLops is still an emerging field, so it can be tempting to write it off as just another tech buzzword, says data science and business analytics consultant Sibanjan Das. But its track record so far has been pretty good, as long as it’s designed in the right way and aimed at the right goal: maximizing model performance and improving ROI. This requires careful coordination between the various components that make up an MLops environment, such as the CI/CD pipeline itself, as well as model serving, versioning, and data monitoring. And don’t forget to create strong security and governance mechanisms to minimize the risk of ML model activities and the possibility of it being compromised.
Although MLops is designed for automation and even autonomy, don’t overlook the human element as a key factor for successful results. A recent Dataiku report noted that over the past year, companies have realized that they cannot scale AI without creating diverse teams that can implement and benefit from the technology. MLops should be a critical component of this strategy because it supports diversification in the development, deployment, and management of AI projects. And judging by Gartner’s MLops framework, a broad skill set will be required to ensure the results deliver maximum value to the enterprise business model.
Even the most advanced technology is of little value if it cannot be successfully passed from the laboratory to the real world. AI is now at the point where it must start to make a valuable contribution to humanity or it will become the digital equivalent of the Edsel: flashy and gadget-packed but with little practical value.
MLops can’t guarantee success, of course, but it can reduce the cost of experimentation and failure, while putting it in the hands of more people who can figure out for themselves how to use it.
The VentureBeat Mission is to be a digital public square for technical decision makers to learn about transformative business technology and transact. Learn more about membership.