Strong gravitational lenses are powerful tools to study the Universe. Lensing is sensitive to the total mass distribution of the lens, allowing it to uniquely probe both dark and luminous matter over a range of scales. In this talk, I present analyses of three separate lenses that each address distinct questions about galaxy evolution and cosmology. Using high-resolution ALMA observations, we can probe the inner mass distribution of galaxies that are lensing a submillimeter-bright source, allowing us to constrain the supermassive black hole at the center of the lens galaxy. I review a pilot study of the lens system SDP.81 and discuss progress on a newly-observed lens, SDP.9, for the purpose of studying relationship between central supermassive black holes and their host galaxies to high redshifts. In addition, strong lensing can be used to constrain substructure in lens galaxies, as well as cosmological parameters. I discuss observations of the Eye of Horus, a double source plane lens (in which two sources at distinct redshifts are lensed by the same galaxy) discovered as part of the HSC SSP, as well as the constraints we will get from follow-up high-resolution observations. From these observations, we can constrain possible substructure in the lens, as well as probe cosmological parameters through the ratio of the Einstein radii to the two sources. Finally, lensed quasars with measured time delays between the multiple images can be used to study the Hubble constant (H0) that is key to probing dark energy, neutrino physics, and the spatial curvature of the Universe. The H0 Lenses In COSMOGRAIL's Wellspring (H0LiCOW) project aims to measure H0 to <3.5% precision from five lensed quasars using deep HST imaging, precise time delay measurements, a measurement of the velocity dispersion of the lens galaxies, and a characterization of the mass distribution along the line of sight. The most recent H0LiCOW analysis of the lens system HE 0435-1223 is able to constrain the time-delay distance to a precision of ~8%. In combination with the previous two H0LiCOW lenses, we constrain H0 to ~3.8% precision for a flat Lambda CDM cosmology, independent of and competitive with the latest distance ladder results. The results are somewhat in tension with the latest Planck results for a similar cosmology, highlighting the importance of this independent probe.