[hospadar]

For us, depends on the seniority of the role, but we've had good luck bringing in people coming from both directions (where I define the "directions" as "software engineering" and "data science/analysts")

Analysts and junior data-science types can often make the transition well if they can beef up their engineering skills (i.e. learn to write tests, make stuff that will be maintained for years)

Software engineers are often a good fit too if they can pick up some of the data skills (get really good at SQL).

Probably really depends a lot on the specifics of the position, sometimes "data engineer" means "write sql queries to apply business rules" and sometimes it means "maintain our interesting in-house ETL applications which were written in Java 8 years ago"

I'd also value the soft skills a lot if I were to be hiring data engineers - so much of the job tends (at least where I am) to be correctly interpreting business rules/needs and anticipating potential future use-cases.

[arithmomachist]

Hmm, so it might be out of reach for me. I have a PhD in pure math, and no experience as a software engineer. I've coded for research, but never for production.