| Dead on that data science job descriptions reveal how poorly companies understand what needs doing on their data and how to ask for it. I don’t agree the answer is to just look for personable, inquisitive people sans PhD and treat them sanely. In the long run few companies are capable of such magnamity and calm. You are typically not doing these nice people any favors, dropping them into a messy pile of data and high expectations. The solution is defining the task you need done and hiring something more specific than a “data scientist”. I have seen Facebook “data science” roles that were just entry level SQL Analyst / dashboard “developer” roles. OK, why not just ask for a “business analyst” or “marketing analyst”? Other “data science” roles are just ETL development jobs; if ETL sounds boring to you, just call this “Software Developer - Data Engineering”. In the rarest of cases, someone justifies a need for an actual ML developer. Ask for an “ML engineer” and pay the sky-high salary. Of course, asking for a DBA is so unsexy that nobody dares do it, and all of the above people - SQL analysts, ETL devs, ML devs - are way less productive than they could be. |