Over the last few years, the advent of the WFC3/IR camera on HST has completely revolutionized galaxy science in the reionization epoch, as it pushed the observational frontier out to z~9-12, only ~450 Myr from the Big Bang. From several large HST programs such as the HUDF09, CANDELS, or CLASH, we were able to identify large samples of more than 200 galaxies at z~7-8, and we are now starting to build up the sample sizes of z~9-11 galaxy candidates. In particular, the recent HUDF12 campaign further increased the depth of the WFC3/IR dataset over the Hubble Ultra-Deep Field (HUDF), and added new F140W filter imaging data. These new data reach to almost 30 AB mag in several NIR filters, which enabled us to detect a sample of nine z>8 galaxy candidates in the HUDF. Using these sources, combined with a search for z~10 galaxies in all the WFC3/IR data around the GOODS-South field, we infer that the cosmic star-formation rate density in galaxies with SFR>0.7Msol/yr decreases very rapidly at z>8, dropping by an order of magnitude from z~8 to z~10. In this talk I will highlight recent progress in exploring the high redshift frontier and in understanding the growth of galaxies in the first two billion years. In particular, I will demonstrate the power of combining deep HST and Spitzer data to directly track the star-formation and mass build-up of galaxies out to the reionization epoch.