Machine Learning System Design Interview, issue 4, Nov 26, 2025
The Infinite Stream Trap
Senior ML Engineer interview at Twitter, and the interviewer asks:
“We have a firehose of tweets coming in at 50k TPS. I need you to maintain a statistically representative sample of exactly 10,000 tweets for a training buffer at all times. The stream never stops. You cannot store the full history.”
Why batch thinking fails in infinite streams - and how Reservoir Sampling saves you
The full answer, with the mechanism and the arithmetic, is for paid subscribers on Substack.