Don't miss Inspire 2024, taking place May 13 - 16, 2024 at the Venetian, Las Vegas. Register Now.


The Two Biggest Challenges in Google Cloud Machine Learning

What's New   |   Bertrand Cariou   |   May 25, 2022

The Competitive Advantage of Google Cloud Platform (GCP) Machine Learning

Among the organizations that have invested in GCP, many are looking to initiate Google cloud machine learning projects. Why? These projects (and machine learning in general) have proved to be a huge competitive factor. In 2017 MIT Technology Review had already claimed that machine learning was “the new proving ground for competitive advantage” and the number of machine learning pilots and implementations has doubled each year since then. Google cloud machine learning projects are paving the way for businesses to understand their customers at a granular level, to detect fraud at its earliest possible onset, to predict product effectiveness and maintenance requirements and much more.


Powering these Google cloud platform machine learning projects is the Google suite, or the most comprehensive AI/ML framework on the market. The Google cloud platform for machine learning provides the necessary storage (such as Google Cloud Storage and BigQuery) and processing (such as Dataflow and Dataproc) for large-scale projects. But where Google cloud machine learning suite excels more than any other cloud providers is with its AI/ML services—BigQuery ML, Cloud AutoML, Cloud TPU, Dialogflow, Cloud Natural Language, Cloud Speech-to-Text and Cloud Translation are just some of the offerings. With the Google cloud machine learning platform, users have the right technology framework to support ML/AI adoption.

The Two Major Challenges to Successful Google Cloud Machine Learning Projects

Even with the right technology, there are still significant challenges to getting these projects off the ground.


  • Low Number of Data Scientists

Given the general demand for data scientists and the high cost of these workers, there’s often a shortage of data scientists available within an organization to take on machine learning. There are a number of solutions for tackling this problem, but what we’ve found most interesting is the concept of the “Citizen Data Scientist” from Gartner. These are technically-oriented resources, though not at the level of a data scientist, who bring unique organizational context to the process. “Citizen data scientists are ‘power users’” Gartner writes that “provide a complementary role to expert data scientists.” With the right technology, such as the Google cloud platform machine learning suite, these types of users can get up and running with machine learning projects.

  • Dirty Data

Any machine learning prediction or recommendation is only as good as the data that feeds its algorithm. As the CEO of Kaggle, an organization that regroups the most talented data scientists in the world, Anthony Goldbloom once said, “There’s the joke that 80 percent of data science is cleaning the data and 20 percent is complaining about cleaning the data.” This is the terrible reality of data science: 80% of the time is spent cleaning data for projects such as machine learning. This might be the biggest problem of all for successful machine learning projects. But luckily for users, Google has a unique solution for the challenge of dirty data: Cloud Dataprep by Trifacta.

Cloud Dataprep: The Secret Weapon of the Google Cloud Machine Learning Platform

Cloud Dataprep by Trifacta is an embedded version of the Trifacta data preparation technology, which has transformed the way that people prepare data. Trifacta guides users toward the right transformation with intelligent recommendations, while its visual interface also allows users to swipe, click, and search their way to understanding and preparing their data—not through code. Therefore, Trifacta and Cloud Dataprep by Trifacta users alike don’t need a laundry list of technical skills to simply prepare data for machine learning models. Rather, Citizen Data Scientists or even analysts have been able to take on this work and are seeing a reduction in time spent preparing data by up to 90%.

Together, the Trifacta technology through Cloud Dataprep and the GCP ecosystem allows users the best of both worlds: the benefits of the Google cloud machine learning platform with the intuitive experience of data preparation with Trifacta. To learn more about improving data quality for machine learning projects, click here.