Open-source contribution
Mesh-Oriented datABase (MOAB)
Contributions to MOAB, the mesh library that underpins how nuclear simulation codes represent and evaluate meshed geometry.
- Role
- Open-source contributor
- Period
- 2024 — present
The problem
Meshed geometry is the lingua franca of particle transport codes: every code that handles CAD-derived geometry eventually needs the same mesh abstraction, and each one used to roll its own.
The approach
MOAB provides that abstraction once — entities, tags, sets, and the file formats that carry them between tools. My contributions went into the parts of the library where the abstraction meets real geometry: correctness of the mesh representation and the tooling that lets users inspect and repair it.
It is the kind of project where the user is often a scientist debugging why a tallied surface is empty. A clear error or a sane default there is worth more than a new feature.
Engineering highlights
The layer everything else stands on
MOAB is invisible until it breaks — and when it breaks, the failure surfaces three tools away from the cause. Work here is defensive by nature.
Repair over rejection
Meshes in the wild are rarely valid. Tools that help a user fix a mesh are worth more than tools that refuse it.
At a glance
- Project
- Mesh-Oriented datABase
- Role
- Contributor
Stack
- C++
- HDF5
- NetCDF