Engineer | AI/ML-accelerated computational science | CI/CD and DevSecOps for research software
AI/ML-accelerated pipelines for experimental and imaging data, from synchrotron tomography and computer vision to multimodal materials measurements and finite-element design optimization, delivered as tested, secure, citable software.
 agentic-bioinspired-cad: text-to-CAD design loop with print, FEA and fracture checks |
 ct-segmentation-toolkit: denoising and segmentation of CT image stacks |
 biomimetic-lattice-pipeline: synchrotron micro-CT of enamel to printed lattices |
 som-multimodal-datareduction: self-organizing maps of multimodal materials data, from J. Mech. Behav. Biomed. Mater. 129, 105147 (2022), Special Issue: New Frontiers in Applications of Artificial Intelligence and Machine Learning in Biomaterials, Organs and Tissues |
| Package | Scope | Archive and release |
|---|---|---|
| ct-segmentation-toolkit | Supervised (U-Net), unsupervised, and label-free (self-organizing map) segmentation of scientific image stacks, with self-supervised Noise2Inverse denoising |   |
| biomimetic-lattice-pipeline | Synchrotron micro-CT of tooth enamel to biomimetic lattices: parametric CAD, finite-element analysis, and closed-loop design optimization |   |
| som-multimodal-datareduction | Self-organizing-map reduction of multimodal materials data (nanomechanics, Raman, fracture) |   |
| agentic-bioinspired-cad | Fully local agentic loop from a text prompt to a certified, single-material, 3D-printable bioinspired architecture, with printability, finite-element, and phase-field fracture checks |  |
Every change to these packages passes one continuous-integration gate: multi-version tests, secret scanning, static analysis, dependency and container vulnerability scans, and a signed software bill of materials. Releases publish to PyPI through Trusted Publishing and to the GitHub Container Registry as signed images with SLSA build provenance, and each version is archived on Zenodo with a DOI.