How AutoML is Killing the Data Scientist

4 min readJan 18, 2022

While working for one Silicon Valley company, I realized the Data Scientist as we know it is a breed dying right before my eyes. Machine Learning has advanced many industries, leading to task automation, better ability to predict outcomes, and deep learning applications to interpret and generate text. Traditionally, machine learning has required specialized data scientists or machine learning engineers to generate value from the available frameworks, AutoML has made machine learning available to the masses. Prior to AutoML, it was common to have data engineers consume and stage application data, analysts to build reporting views from these data sources, data scientists to perform feature engineering and model tuning, and machine learning engineers to productional the inference pipeline. With the availability of AutoML, business users such as analysts, data engineers, or even technically-minded business users can reap the value that machine learning provides. The benefit of this is a significant reduction in cost and the required infrastructure to support such advanced analytics platforms. With data scientists free to spend their time on more productive matters for their skill sets, AutoML will enable the advancement of machine learning capabilities, which can further the field of data science.

In traditional data science applications, entire teams of engineers were necessary to support advanced analytics capabilities, such as machine learning. With the advancement of AutoML, the amount of specialized knowledge and resources is dramatically reduced. For example, a professional such as a data engineer could identify and curate a data set, design a feature set, train a model, and output results to a data warehouse, and bring the production-sized output to the business users that will benefit from the predicted results. With this invention, organizations with resource constraints who may have otherwise been unable to prioritize machine learning can now benefit. While machine learning engineers are some of the highest-paid titles in software engineering, their cost can be many times justified by their specialized skill set of leveraging AutoML applications to produce results that the organization can use to make strategic decisions.

Prior to AutoML, whenever an organization would look to build predictive models, the individual identified would likely be a data scientist, with a background in a deep mathematical field. In this model, someone like a PhD statistician would be employed to build advanced…


Data Engineering enthusiast, mentor, data geek, passionate about great technology and process