Machine Learning System Design Interview #40 - The Look-Ahead Trap
Why predicting the past with future logs quietly destroys your live performance, and how lagging your training features fixes the hidden pipeline latency killing your metrics.
You’re in a Machine Learning Engineer interview at Netflix and the interviewer asks:
“You train a predictive model on user activity logs using a standard random 80/20 split and hit a spectacular 98% offline accuracy. But the minute you push it to production, online performance crashes to 55%. What structurally broke, and how do you fix it?”
Most candidat…
