Forge Range · Aurora Forge Lab
0 / 35
35-day range · built for 2026

Interviews stopped testing recall.
They test how you drive the AI.

A month of exercises for the human-led, AI-assisted coding round — the format Meta is piloting and rolling out, and that Google has reportedly tested for selected SWE interviews. Read unfamiliar code, scope a vague ticket, validate what the model wrote, and ship. Every day is a runnable test suite and a spoiler-free mock.

35days
7weeks
L3–L5difficulty
hours of reps