At Oreyeon LDA, I lead the AI infrastructure function for runway safety, building cloud systems across AWS Commercial and GovCloud. I developed secure environments with controls supporting CMMC 2.0 Level 2 and ISO 27001 requirements.
I engineered over-the-air model deployment for runway vehicles, including ONNX conversion, encryption, and device-specific TensorRT compilation. I also designed MySQL databases for deployed-model metadata and performance metrics, and deployed an authenticated MLflow server for experiment tracking.
I led development of Polaris, Oreyeon LDA’s internal data management, governance, and analytics platform. I implemented image tiering with projected annual AWS savings of approximately $48,000, and developed image profiling, validation, clustering, and deduplication workflows.
Earlier at Oreyeon LDA, I improved a PyTorch classifier to 94% top-1 and 99% top-3 F1 scores on a held-out test set, and optimized models for mobile deployment using quantization. I’m also participating in SPAR research on AI-related cyber risks to low Earth orbit satellites, applying the STPA-Sec framework.
