Helping data roles understand each other
Why do data engineers, analysts, managers, and ML people so often seem to talk past each other? This talk looks at the real, painful, and sometimes funny ways how misunderstanding between data roles happens and what we as data folks can do to make working together less annoying.
If you ever worked in a company with decently sized data department, you probably noticed how sometimes communication between teams is the main pain point.
Data engineers, analysts, managers, machine learning specialists, and other data roles often look at the same problem from very different angles. While often this is useful, it also backfires quite often and leads to confusion, frustration, and borderline animosity. And as work gets more and more automated, the interhuman communication becomes more and more important
I have worked in different teams and in different roles, and in this talk why this miscommunication happens, what fosters it, how each role gets shaped by its daily work, and how we can address this to make these relationships at least less painful and maybe even enjoyable.
This talk will use a mix of role "avatars", and I plan some audience interaction in style of "8 out of 10 cats", and there would be a fictional (but very realistic and inspired by real events) data team story to show how things tend to go wrong. Then I'll look into practical ways to improve communication, set better expectations and understand what other data people are actually trying to do.
And yes, of course, I would say a few words about AI and stuff.
After doing two PhDs, one in Theoretical Physics and one in Materials Science, I decided that maybe there was such a thing as doing too much science.
As a recovering academic, these days I work across various data roles, doing a bit of everything: engineering, modelling, machine learning, and analytics. During my time in computational research, I picked up a few practical tricks that turned out to be just as useful in industry. I enjoy sharing them because I’ve found that some of them are not as widely known in the Python and data community as I expected.