Table of ContentsIntroductionWhat exactly are you trying to improve?The fantasy of perfect communicationDo not start with valuesThe system benefits from its current stateCase studies: the downsides of transparencyThe five types of interventionsEvaluating the effectiveness of interventionsWhat are you willing to give? IntroductionYou may be a CEO, a department head or a team lead. Somehow you
Data Scientist: Why am I not receiving a bonus this year? I am constantly coming up with new ideas, putting my heart into it.Manager: It is all based on results. This year your research has not produced outstanding results.Data Scientist: Research is inherently random. It may produce something outstanding once in five years.Manager: Then you
Manager: I want a model that predicts X. Data Scientist: What accuracy is sufficient?Manager: It has to capture whatever signal there is in the data.Data Scientist: Let me see… How will the model be used?Manager: We will decide later. Let’s have a good, working model first.Data Scientist: What is the cost if it makes a mistake?
Manager: I have told you to start doing research on Z last month.Data Scientist: Yes, you did.Manager: It was your only priority that you have been working on, hasn’t it?Data Scientist: Yes, it has.Manager: When will it be ready for deployment?Data Scientist: In about 2 months.Manager: But last month you also told me about 2
Manager: Last week you told me that the model was ready.Data Scientist: Yes, but on Tuesday I found a bug. Column X in the data had a lookahead bias due to some quirks. So now I am retraining the model. The existing version is complete junk and should not be trusted. But we’ll get the
Page [tcb_pagination_current_page] of [tcb_pagination_total_pages]