The Matter Compiler will take a device design and produce a physical part without the manual translation steps. Design for manufacturing (DFM) is where that translation happens, and it is currently human work: an engineer reasons about how a part must be arranged, held, and processed, and encodes that judgment one design at a time.
This role owns that reasoning as software. It is the DFM function inside the Atomic Machines CAM stack, the engineering discipline of turning device geometry into manufacturable geometry under real process constraints, generally rather than case by case.
The scope is the full DFM layer: the geometry between a device model, the workpiece, and the machine processes, including how parts are arranged on a blank and held in place during cutting; the DFM rules for each process and material the platform supports; the constraints and checks that tell a designer a part cannot be made as drawn; and the physical models that ground those rules in what the processes do to the part.
The engineer in this role is the person on the team who thinks in manufacturing constraints and writes code that respects them. Design engineers bring geometry that cannot yet be built. Process engineers bring results from the machine that the rules did not predict. The person in this role connects those two and works inside a cross-functional team spanning AI, Modeling and Simulation, Design, and Process Engineering.
- DFM as a software capability. The algorithms, representations, and constraints that convert device geometry into geometry a process can execute. This covers part arrangement on a blank, retention during processing, and release afterward, and it expands as we add processes.
- Manufacturability constraints in the design loop. Encoding what our processes can and cannot do, so infeasibility surfaces at design time rather than at the machine.
- The bridge between process intuition and code. Working directly with design and process engineers to elicit the judgment they apply by hand, formalize it, and make it auditable and testable.
- Physical grounding. Moving DFM decisions from heuristics toward criteria based on the mechanics of the process, with our Modeling and Simulation team.
- Validation against reality. Defining what correct means for a layout, testing against fab runs, and folding failures back into the constraints and models.
- The knowledge base. Turning our production history into a structured record that supports calibration, regression testing, and eventually learned components.
- This posting is not tied to a specific level and spans early career through Staff, or L4 to L6. Candidates should have a minimum of 5 years of relevant industry experience or a PhD in a related field.
- Practical DFM experience, demonstrated by work where you wrote code that generates geometry under real manufacturing constraints. Relevant examples include slicer or toolpath software for additive manufacturing, non-standard toolpathing strategies such as continuously self-supporting structures, design software for sheet metal stamping or other tool and die applications, PCB or lead frame layout, or comparable design automation work where geometry is constrained by physics rather than by convention.
- Working computational geometry ability: 2D boolean operations, polygon offsetting, packing and no-fit-polygon style reasoning.
- Strong software engineering: Python plus a systems language, and comfort driving geometry kernels and libraries through their APIs (Shapely, Clipper, OpenCascade, CGAL, or similar).
- A clear demonstration of working productively on novel, poorly specified problems. A PhD is one way to show this. Open source contributions, patents, or industry work on greenfield problems count equally.
- Willingness to ground your work in physical evidence from the fab, and to iterate with the engineers running the process.
- Bachelor's, Master's, or PhD in Mechanical Engineering, Computer Science, Applied Math, Computational Design, or a related field.
- Exposure to laser micromachining or other subtractive micro-scale processes: kerf, heat-affected zone, tabbing, part release.
- Enough mechanics background to reason about part stability during processing, or the interest to build that with our Modeling and Simulation team.
- Combinatorial and geometric optimization, using MILP, constraint programming, or metaheuristics.
- Machine learning on geometric data, for example learned models over meshes or B-rep graphs, neural fields, or learning from expert demonstration. This is a growth direction for the role, not an entry requirement.
- Experience placing heuristic or learned components inside a deterministic, auditable pipeline, including validation and fallback behavior.
- Familiarity with CAE tools (e.g., Comsol, Ansys, Abaqus).
- Contributions to open-source geometry or manufacturing software.
The compensation for this position also includes equity and benefits.
Skills Required
- At least 5 years of relevant industry experience or a PhD in a related field
- Practical design-for-manufacturing experience writing code that generates geometry under real manufacturing constraints
- Working knowledge of computational geometry, including 2D boolean operations, polygon offsetting, packing, and no-fit-polygon reasoning
- Strong software engineering skills using Python and a systems programming language
- Experience working with geometry kernels or libraries through APIs, such as Shapely, Clipper, OpenCascade, or CGAL
- Demonstrated ability to work productively on novel, poorly specified problems through a PhD, open-source contributions, patents, or greenfield industry work
- Willingness to ground software development in physical evidence from fabrication and collaborate with process engineers
- Bachelor's, Master's, or PhD in Mechanical Engineering, Computer Science, Applied Mathematics, Computational Design, or a related field
- Exposure to laser micromachining or other subtractive microscale processes
- Mechanics knowledge related to part stability during processing
- Experience with combinatorial or geometric optimization, MILP, constraint programming, or metaheuristics
- Machine learning on geometric data, including meshes, B-rep graphs, neural fields, or expert demonstration
- Experience integrating heuristic or learned components into deterministic, auditable pipelines with validation and fallback behavior
- Familiarity with CAE tools such as Comsol, Ansys, or Abaqus
- Contributions to open-source geometry or manufacturing software
What We Do
Atomic Machines is redefining humanity’s relationship with matter. We see a future where our tools will allow us to reorganize matter at the atomic level at will, where we will go from bits to atoms for any object or machine that can be designed in alignment with physical laws. We have begun our journey with the development of a robotic manufacturing platform capable of making an entirely new class of micro-electromechanical (MEMS) devices. We are well funded and have exceptionally strong product/market fit and a clear go-to-market path for the device we will make first with our platform. Our platform breaks traditional manufacturing paradigms and constraints, enabling inexpensive rapid prototyping as well as large scale manufacturing with highly compelling economics. Joining forces with us means becoming part of an incredibly talented, inventive and passionate multi-disciplinary team working on a massive world-changing mission. You will have the opportunity to help define the company from its early days. You’ll be challenged to learn and grow as a builder and a leader as the company itself grows rapidly. And you will receive significant equity compensation - you’ll truly be a company owner and benefit financially from our overall success.
