Machine Learning-Enhanced Spectroscopy Technology Service

The buyer requires a machine learning-enhanced spectroscopy technology service for high-resolution radiation detection using low-cost detectors. The requested scope includes deploying a data-driven, convolutional neural network to reconstruct high-resolution spectra from low-resolution inputs. This 1.6M-parameter model uses 6.2 MB of memory to enable rapid inference on low-power devices. The detector-agnostic system adapts to gamma, x-ray, neutron, or charged-particle spectroscopy, and scales to NaI, CsI, or plastic scintillators. This setup achieves energy peak discrimination comparable to high-purity germanium detectors without requiring cryogenic operation or liquid nitrogen, making it suitable for remote, field, and mobile environments. Only onshore United States organizations are eligible to apply, and the selected vendor will perform all work offsite. The buyer is currently looking for proposals and has not provided specific budget figures or key dates.