I just stumbled on this thread by accident, having built a prototype UV enlarger myself over the past weeks for cyanotype. I use it to enlarge my 6x7 negatives to A3 paper size (11-3/4 x 16-1/2inch, 297 x 420mm)
The key to this game isn't using as much light as possible, but making the enlarger efficient in terms of its light path and its transmission. That means using acrylic fresnel lenses for the condensor, and setting the whole thing up as a point source enlarger. For point source enlargers, the light source should be smaller that the exit pupil of the enlarger lens, otherwise the light cannot pass the lens. So using the latest high power UV LEDs with a quarz lens is the best approach. the dies in the LED I use measure 7x7mm, this allows me to use a regular 100/5.6 enlarger lens.
Using a LED with a larger cross-section is not effective; you'd need a faster lens to pass all the light. There are not many fast enlarging lenses, and fast lenses that do exist are either poorly corrected for near UV, leading to massive focus shift, or optically inferior for the application due to field curvature or poor sharpness for other reasons.
I also found that for the cyanotype process a 385nm wavelength works very well. There is a trade-off between transmission of the enlarger lens and sensitivity of the cyanotype process. Below 380nm the transmission of enlarger lenses falls off steeply. The spectral sensitivity of cyanotype really drops off at about 390nm. So there is an optimum based on the two factors.
Focusing isn't really an issue. There is a focus shift also with enlarging lenses that are corrected to near-UV (they pretty much all are), so a grain focuser is not very useful. I use copier paper on my easel. The optical brighteners fluoresce blue as a result of UV illumination. With a pair of UV safety glasses you can focus by eye. If you print the same format all the time, you can fine-tune with a piece of paper placed on a wedge.