I might work differently from others but I do an initial (light) sharpening along with any global changes (temp, color, brightness, exposure, etc) before I do any pixel level changes. After pixel level changes, I do snapshot then slight area sharpen (USM with tiny radius, 0 threshold). Then save as PSD.
If I am going to print I size and do flnal area sharpen, smart sharpen, convert profile, convert to 8 bit, save as jpeg.
If I am going to display on web, I reduce size, convert profile to sRGB, convert to 8 bit, resharpen and save as jpeg