Help us turn planning into something faster.

Planlift is looking for a founding CTO who wants to build something from scratch rather than inherit someone else's architecture. You would be the first technical decision-maker at the company, working directly with the founding team to set the direction of the product and the engineering culture that grows around it.

The problem is a real one. Architects, planners and developers spend days creating presentation visuals for projects that are still in early concept. We compress that into seconds. You would own the systems that make that possible, from image generation pipeline to the web platform professionals use daily.

This fits someone who has been in a senior engineering role and wants ownership they have not had yet, or a technical co-founder who knows exactly what they would do differently this time.

This is a hands-on role for a senior engineer who wants to build something close to zero. You would be among the first engineering hires, working with the CTO to ship the core product and set the technical patterns the team will scale into.

The product has real surface area: file uploads, rendering queues, credit systems, user accounts, a generation interface, and a sharing layer. It needs to be fast, reliable and clear under real professional use. That requires engineering judgment, not just feature execution.

If you want to move into pure management in the next year, this is probably not the right fit. If you want to be the person who set the architectural tone and can still point to the commits five years later, keep reading.

Planlift takes a 2D drawing and returns a photorealistic 3D visual. The magic in the middle is what this role is about. We automatically write the generation prompt from the uploaded drawing, handle conditioning on the drawing's structure, and tune the output for architectural and planning contexts specifically. You would own that system.

This role is for an ML engineer who is drawn to hard applied problems rather than benchmark competitions. Our outputs need to be good enough for professional use, which means they have to look correct, read as plausible architecture, and remain faithful to the original drawing. That is a more constrained problem than open image generation, and a more interesting one.

The design and planning expertise is at the table. What we need is someone who can translate that into models that work reliably in production, not just in notebooks.