Curriculum

Data fluency

The report names SQL as the single most-named tool (27), ahead of APIs (24) and MCP (23), and flags that no course teaches it: this is called out twice, in the missing-material list and as one of the sharpest JD-versus-course disagreements

1. 02.01

   SQL for PMs

   You write the five queries a PM actually needs, a filtered SELECT, a GROUP BY, a JOIN and a date range, and read the answer yourself instead of waiting on the data team.
2. 02.02

   Read the logs

   You mine conversation logs, support tickets and traces for intent in the user's own words and for the failure samples everything downstream is built from.
3. 02.03

   Cohorts

   You define the cohort window and the denominator behind a rate yourself, and separate a metric you can defend to leadership from a vanity number that quietly hides a mix shift.
4. 02.04

   From raw data to a golden sample

   You turn raw logs and query results into a curated golden sample, sampling failures on-distribution so the eval set the later chapters consume measures the errors that actually happen.
5. 02.05

   INTEGRATION CASE

   Build a counted table of failure types from real runs, so you can sort fixes by frequency, severity, and evidence.