A general 1D-Var retrieval infrastructure has been developed with the design philosophy based on emphasizing the transition of basic research into the operational stream. The infrastructure covers each of the key components for a successful retrieval/assimilation. A fast and accurate forward radiative transfer model is essential for a robust retrieval. OSS represents the state of the art of fast RT models, under both clear and cloudy conditions, and has already been successfully incorporated into both the Community Radiative Transfer Model (CRTM) and is part of the NPOESS operational system. It is also essential to that the fast RT model is trained and validated with the most up to date LBL models. LBLRTM is the state of the art in LBL models, incorporating the most up to date line parameters and line shapes. The general infrastructure has allowed recent developments in cloud and surface parameter characterization/models to be incorporated into the general retrieval product generation routines. Retrieval technology, developed to optimize the NPOESS CrIMSS/CMIS retrievals, have been incorporated. Infrastructure has also been developed for post post-processing and validating retrieval products. The validation methods are applied operationally at AER for the data collected with the NOAA AMSU instruments. This will serve as a model for the NPOESS CrIS validation for which the METOP IASI can be used as a prototype.