Near-infrared (NIR) spectroscopy is the rapid, workhorse technique used by recycling facilities for sorting of plastics, yet as currently used it is insensitive to variations in molecular architecture reducing its use for sorting chemically similar polymers such as high-density polyethylene, low-density polyethylene, linear low-density polyethylene and polypropylene. To overcome this challenge, we have previously shown that utilizing machine learning, real-time NIR measurements can be correlated to molecular architecture sensitive properties, such as short chain branching and degree of crystallinity extracted from much slower, orthogonal techniques including multidetector gel permeation chromatography, differential scanning calorimetry, and density measurements for idealized polyolefins [Macromolecules?2024, 57, 5, 2329–2338].  Proposals will build on the prior work by exploring more complicated systems such as post-consumer resins, property changes from physical or chemical processing, the addition of other measurement modalities and/or using more sophisticated machine learning techniques. The proposal will emphasize the role of metrology, provide findable, accessible, interoperable and reuseable (FAIR) data sets, seek to improve sortation technology for homogeneous recycling feedstocks and work to advance the US circular economy.