Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
There's a position being adopted at many companies called "data science engineer." Their job is to bridge that gap, and provide the engineering discipline, tools, infra, and general connective tissue to hook data science work into company systems effectively.
It's a pretty good setup all things considered.
1. Study the codebase, not just the database. You usually don't have to be an expert at the backend language to just read its logic. I feel like most data scientists lack an appreciation for how data was sourced. Pretty much any company has legacy stuff and badly designed databases, so without a global understanding of the application logic you will probably make wrong assumptions.
2. For engineers, one of their biggest nightmares is a data scientist who asks "if you build that feature, could you add these 4 extra columns to the database to record these statistics". They should only want data that's directly relevant for the application in their data storage — so help them by proposing a solution which allows them to keep their DB as a highly normalized, single source of truth. Having a single source of truth eventually helps data science as well! If you need extra data which can't be derived from existing application data, set up a separate system to log these.
mrtehseen32418dI am a full stack developer turned Data Scientist and believe me I still get stuck with some integrations with organizations, there should be a line between jobs, they overlap but in a safe ecosystem it shouldn't be