sizeless is building the data layer for the infrastructure under our streets. A construction crew films an open trench with a phone, and our system turns that recording into a 3D model, recognises what is in the ground, and produces the documentation the industry needs - CAD, BIM, GIS and quantities - in a fraction of the time it takes today.

It is a real problem with real customers: construction and utilities still document reality by hand, and almost nothing about that process has changed in decades. We are a small technical team, founded out of ETH Zürich and MIT, and we are looking for engineers who want to work on frontier AI in a domain where it has barely arrived.

Tasks

• Build and improve the agents that turn captured reality into finished documentation
• Work on the perception and semantic layer that recognises components and structures in 3D
• Build the evaluation and verification that tells us when the system is right - and when it is not
• Scale the pipelines behind it: video, point clouds, large documents, and a lot of compute

Requirements

• Solid engineering fundamentals and strong Python
• Hands-on experience with LLMs, agents and the tooling around them
• Real leverage from AI-assisted development, whatever your setup
• Ownership: you take problems personally and see them through
• Full working proficiency in English; German is a plus, not a requirement