High-redshift radio galaxies (HzRGs) are vital laboratories for studying massive galaxy formation and evolution in the early Universe (see review by Miley & De Breuck, 2008, A&ARv, 15, 67). We have developed a new selection technique for finding these rare, powerful systems at z > 5, making use of spectral curvature in the Murchison Widefield Array (MWA) 70-230 MHz GLEAM survey (Hurley-Walker et al., 2017, MNRAS, 464, 1146). In a pilot study of four sources, we have discovered the second-most distant radio galaxy currently known (z=5.55; Drouart et al. 2020, PASA, 37, e026), with the exciting possibility that a second source is also at high redshift (currently being followed up with approved HST time). In this talk, I will give an overview of our pilot study, as well as ongoing work to expand our technique to a new sample of 90 HzRG candidates in the ESO VIKING near-infrared survey area. Our goal is to build a sample of Type 2 radio-loud active galactic nuclei at z > 6.5 during the Epoch of Reionization, thereby allowing us to study the co-evolution of the supermassive black hole and host galaxy during the first billion years of the Universe. This is very much a multi-wavelength project, making use of data from a variety of telescopes (e.g. MWA, VISTA, ALMA, VLT, and ATCA), and I will conclude with some thoughts on the scientific possibilities and synergies offered by the next generation of multi-wavelength facilities for the field of HzRG research.