Rant: 'Director of Data & AI' can't explain his own realtime pipeline

teodorio · x · 2026-09-30

The author recounts a painful conversation with a company's 'director of data and AI.' Asked how Redshift serves production traffic given its known scaling and latency limits, the director could only repeat 'we split the Kinesis stream, one goes to Redshift, one to Lambda.' Pressed on why aggregations run in Lambda instead of Flink and where results are stored for the frontend, he eventually replied, 'uhhh who are you again?'

The author calls it the biggest waste of time and swears off such events.

Related event: Developers Call Out Vague AI and Data Practices as Experts Fail Basic Questions(2 posts)→

Original post →

More from Fun

Fun channel →