Snapchat Machine Learning Engineer Interview Experience — Onsite and Team Matching

My onsite covered ML fundamentals, coding, system design, and applied ML. Fundamentals included optimizers, normalized entropy and other evaluation metrics, confusion-matrix terminology, calibration, and overfitting.

The coding task involved finding the peak of an array with increasing and decreasing sections, followed by a discussion of how equal values changed the search assumptions. Both design discussions involved video recommendations. Topics included data flow, two-tower model updates, labels and features from logs, serving, and MMOE targets and losses.

After the debrief, another system-design round was added on identifying harmful content. I passed and moved to team matching.