Hey, thanks for considering a role at Virtual Staging AI!
Below you’ll find a problem statement which is relevant to our work.
It is often the case that members of our team have to tackle a novel problem with many potential solutions, so our goal here is to see how you break down a big project into pieces and use your intuition to prune the solution tree and quickly get to a prototype. It’s also important for us to see you think about potential improvements and experiments which are outside of the scope of this challenge.
We try to define interesting and stimulating tasks, but are open to changing them! Let us know if you want to take this somewhere else, or see adjacent tasks that interest you more!
Please don’t spend more than 7h in total on the task. Feel free to split up the time if it better suits your calendar.
Example of virtual staging
Virtual staging is the process of adding virtual furniture to a photograph of an empty room in a way that is consistent with the layout, lighting, materials, and stylistic attributes of the room.
The goal is to preserve the existing room, especially structural elements like windows and doors, while making the added furniture look as appealing and realistic as possible.
Real estate agents and photographers use this service to showcase the potential of a home to potential buyers. The goal of photorealism here is not to fool buyers into thinking the furniture is real, but to generate appealing visuals of what the space might look like once furnished.
Until last year, virtual staging was almost exclusively performed manually by designers and VFX professionals.
Implement a proof of concept for virtual staging using segmentation-guided diffusion models.
Scope the problem down as needed to have something runnable within the given time-frame.
Input
Output
Please provide: