Python doesn't make a data scientist
What makes someone an expert is their mental models
What are you looking for when you hire a specialist like a biologist, a chemist, a software engineer or data scientist? The job description may list specific tools and experience such as programming languages or data types. But if all they’ve done is memorized Python syntax or copied some tutorials using the kind of data you care about, they’re not going to get the job done.
What you’re actually looking for is their ability to address problems in their particular domain. It’s not Python syntax that matters; it’s knowing how to approach the kinds of problems that you can solve with it. It’s not the specific ways they used data in the past that matters; it’s knowing how the quirks and complexities of the data type will determine the solution to your new problems. The experience you list on the job description is a proxy for the problem solving skills you actually want.
In other words, what you’re really looking for is a well developed task mental model: how they frame and prioritize goals and problems, how well they understand the tools available and how they evaluate which tools to choose in specific circumstance. If they have that, they can Google the rest.
Mental models are what make each member of your organization effective as an individual. They also determine how these individuals come together as a team. If you hire great people, they’ll bring great mental models for their individual tasks. It’s up to the organization to help them adjust these models to work as a team.
Jesse: programming language has no correlation with anything around working with data in any domain. Python or C# or Java or whatever are all fine. As you state, and from a vendor standpoint this is our design challenge, is to find generally applicable ways to take data collected in experiments and map it into some useful analytic or ML or DL such that it can serve as an adjunct to the scientists in discovery or data interpretation/understanding. At Sapio we are starting initially with Flow Cytometry data ( https://www.sapiosciences.com/flow-cytometry-data-analysis ), but there is so much more we can and will be doing in life sciences domain. It's exciting time having all these new methods and approaches that could be brought to bare on tough data problems.