If it has to be gold - it's going to be very hard. Processes that leave a single atomic layer rely on self limiting processes where the substrate (gold powder in your case) catalyzes a chemical reaction on the bare surface, but not on the areas that are already coated (such is the growth of grphene on copper). And gold is particularly bad at this because it is a noble metal, therefore a poor catalyst.
Other deposition methods, such as MBE (molecular beam epitaxy, works by sublimating material by e-beam or resistive heating) or PLD (pulsed laser deposition, where a powerful laser pulse ablates small amounts of matter onto a substrate) or sputtering (eroding material from a "target" source onto a "substrate" using bombardment by energetic particles), can deposit atoms slowly enough that a film with an average thickness of one atom can be reached. Whether the film can be heated to coat the surface uniformly depends on the chemistry and interactions between the substrate and the deposited species (basically on the surface tension). Films that grow in the Volmer–Weber mode will form islands and never a uniform film.
All of the above refers to continuous, single crystal substrate. Growing things on powder is much worse, because heating the powder to increase the mobility of the coating atoms to make the film uniform will allow them to diffuse between the gold grains and into the "bulk" of the powder cube.