In-line wafer-map watchdog at the equipment edge: zero-copy SECS-II decode, CPU/GPU spatial-signature detectors, FabEye's CNN on a custom CUDA inference engine, and a lot hold sent back to the tool in under a millisecond.

Wafer-map pattern recognition on real WM-811K data, with lot-disjoint evaluation, conformal prediction that holds on unseen lots, and an ONNX + FastAPI service.

Fab yield excursion detection on a simulated fab with ground truth: SPC and root cause scored against it, wafer-map patterns from FabEye, real-time SPC on Kafka, Grafana and Power BI.

Deep learning platform for inverse molecular design — start with a desired property, generate the chemical structure that satisfies it.

High-performance deep learning framework for thermodynamic phase identification from molecular dynamics simulations.