We are seeking a Geometry Processing developer to research, develop, and maintain a set of internal libraries for geometry processing. You will play a pivotal role in developing cutting-edge technology that is capable of automatically generating engineering designs that conform to manufacturing, physical, and aesthetic considerations.
Braid works on engineering design automation. Our view is that producing a good design means reasoning about physics, geometry, and manufacturing constraints together, from first principles, rather than from patterns learned by past examples. This is a problem that more data and larger models do not solve on their own: it needs a combination of formal reasoning and learning, and people comfortable with both.
Our reasoning architecture is built to replace the draw→simulate→redesign loop. An engineer describes what a component has to do and our reasoning agent derives geometry that meets those requirements and can be manufactured, in hours rather than weeks.
See more on our website: https://braid.tech/
Our career pages: https://braid.tech/careers
Geometry is where AURA's reasoning turns into physical form. Everything the system proposes must ultimately resolve into valid geometry that meets real physical and manufacturing constraints. This role owns that core: building the geometric representations and algorithms that make it possible.
We do not lock ourselves into a single approach. Depending on the problem, the best solution might come from differential geometry, combinatorial algorithms, or numerical optimization. You should have enough range across geometry, graphics, and numerical methods to pick or build the right tool for the job.
This is a practical software engineering role. We are building production software for real users, so algorithms must be fast, robust, and able to handle noisy input. The focus is on shipping reliable code, not writing research papers.
Own core geometry algorithms and data structures. Develop the representations and algorithms that transform system logic into reliable physical geometry that engineers can depend on.
Apply the most effective methods for each problem. Draw from differential geometry, computational geometry, graphics, and numerical optimization to select or design the best approach for complex problems.
Integrate real-world engineering constraints. Build geometric solvers and workflows that inherently respect physical and manufacturing limitations throughout design and generation.
Build robust software for noisy, real-world data. Deliver production-grade geometry routines that run quickly, handle imperfect input gracefully, and scale reliably in a production system.
We look for a particular combination of qualities.
Depth in one area, and range beyond it. The work sits at intersections like geometry, physics, manufacturing, and machine learning, held together by systems that have to be correct and fast. We look for people with real depth in at least one of these who are willing to learn enough of the others to be useful where they meet. We are wary of specialists who do their own part well and leave the seams to someone else.
Driven to make the product better. Braid already has a working product, users, and an architecture that enables production runs. Much of the work is making that better: strengthening it, simplifying it, and extending it. We look for people who find that satisfying, and who would rather take something concrete and make it excellent than start from a blank page for its own sake.