I interviewed for a Staff-level engineering role at Luma AI. The recruiter screen was positive and I was moved to a take-home assessment. The assignment was not a small coding exercise. It required building a complete end-to-end product, making it deployable/publicly accessible, writing documentation, recording a video walkthrough, and submitting AI session history. It was estimated at 8–12 hours, which is essentially a full unpaid working day if done seriously. The technical challenge itself was interesting, but the communication around the process was poor. Role alignment was unclear, follow-up was slow, and after submitting the assignment I received a generic rejection with no specific feedback. I then followed up politely asking for brief feedback, especially given the size of the assignment, and received no response. The overall experience felt one-sided: Luma asks candidates for a significant amount of unpaid work, but does not provide meaningful communication or feedback in return. Candidates should understand that they may spend a full day building, deploying, documenting, and recording a project, then receive only a template rejection or no useful follow-up.