@Enginerding: Aw maaaaan.... But I'm so comfy here! Hehe. Sounds great, thank you for lending your medical experience!
On the UVC side of things, if I'm reading the numbers right from the study, it might be completely feasible to stick an off the shelf UVC lamp in an opaque can and pipe exhaled air through it without much need of anything else. Geometries and seals will be important, but need to figure out the UVC source first. Even if the can is 3d printed or made from chunks of PVC pipe, the plastic should probably hold up against photo degradation long enough for short term use. I haven't found any good source of info on how long it takes UVC to kill plastic. Also, from looking at other info, 254nm UV-C apparently eliminates ozone and does not actually create it (lower nm UV generates ozone). I'm currently trying to find UVC lamps that might work for an in-line virus killer for CPAP/BiPAP/etc exhaled air exhaust.
Let's say we have a system of two bags where each bag can contain the exhaled air of 15min ventilation (that's roughly 20l). When a bag is full the UVC source kicks in for 15min, disinfecting the contaminated air while the other bag fills up. We could either use fully collapsable bags to use their full capacity (bag gets squeezed->no remaining air) or we have to apply some sort of pressure to get the 20l of air into a container that is already filled with air at ambient pressure (having a pressurized container of airborne viruses kinda sounds like a bad idea).
A 20l cylinder (20cm diameter, 65cm height) has a surface area of 0.4m^2 (or 4000cm^2), that means we would need a 16W UVC source if we assume that the intensity for UV radiation of water samples is also viable for an aerosol (probably not). Yeah, that could work. I'd use a 50W source just to be on the safe side and some bigger containers to minimize the required pressure (a 20l canister would sit at 2bar when loaded with the 20l of residual air+ 20l of exhaled air), but it sounds feasible.
Here goes my current thought:
(Potentially dangerous) Assumption #1: The study only used a single intensity level "4016 μW/cm2 (where μW = 10−6 J/s)" which resulted in "400-fold decrease in infectious virus" in 6 min, and total eradication to detectable limit in 15min. There is no data on the relationship between UVC intensity and required exposure time in this specific study. If someone knows, please provide info!

Using the study's exposure intensity as the MINIMUM and if we (potentially dangerously) assume the relationship between intensity and required exposure time is somewhat linear, then increasing UVC intensity to somewhere around 900x (into the W/cm2 range, which is still do-able), the required exposure time would reduce to around 1 second compared to 15 minutes.
(Potentially dangerous) Assumption #2: The whole "social distancing" recommendation of 6ft/2 meters (please correct if better rec's available) is because of gravity's affect on virus transmission. If the output port of a virus killing device is piped down to the floor where a mop and bucket can periodically finish the job, then the targeted virus killing result could potentially get moved to a less conservative ~99% instead of 100% reduction in detectable virus infectivity. So, if we went with this (potentially dangerous) assumption, then the targeted exposure time required for the lower intensity UVC used in the study reduces to their 6 minute value.
(Potentially dangerous) Combination of both assumptions: You might be able to make an in-line virus killer that maintains enough UVC intensity for the designed geometry of the "can" to kill viruses at or above the peak air-flow rate of a patient's exhale. Basically, kill the suckers in real-time. You probably won't kill every single one of them, but depositing a few active viruses under a massive pile of dead virus carcasses on the floor is a major reduction in risk for anyone around a patient, whether at a hospital or at home.
Big open ended question of the unknown: Is UVC
more/less/same as effective on
water/air/aerosol/surface dwelling viruses?