I feel like in the near future there will be a more formal hybrid role between data engineering and data science - like devops or full stack developers. The best data scientists I have worked with (ML mostly) have been incredible data engineers as well - some of them former sysadmins, backend developers or DBAs themselves. They know where to get the data, how to set up pipelines and jobs, how to make sure they run properly, best practices for reading and writing to databases so they don't fall over, error reporting and logging, hosting their inference models on APIs, security by design... The amount of back-and-forth that gets cut out to go from raw data to product is huge when you compare it to a traditional siloed setup of business analysis/stakeholder, data science, infrastructure/security and systems engineering.
I know some people cringe (mostly infra) when they think of data scientists having direct access to databases and infrastructure but honestly you should have a level of understanding and responsibility to get there.
The data scientists that do data engineering are usually much more valuable to the company and definitely earn more.
I agree, up to a point. I feel that companies up to a few billion dollars in stable revenues in a non data intensive business don't need teams of Data Scientists, but would benefit from some people having Data Science skills (both Data Engineers and Business Analysts).
I know some people cringe (mostly infra) when they think of data scientists having direct access to databases and infrastructure but honestly you should have a level of understanding and responsibility to get there.
The data scientists that do data engineering are usually much more valuable to the company and definitely earn more.