> do the artificial structures humans use, like house frames or bridges' trusses recover strength:weight ratio?

There is some improvement from computer design. We have better tools to know the margins that allow us to remove as much material as possible, things like that. So we can tweak existing designs to make them more efficient, like adjusting the section of beams to make them as strong but with a few percent less materials.

But the major improvements come from using different materials and new techniques, side-stepping the scaling issue.

For example, we could not scale old arch bridge designs very far. Getting steel beams let us build truss bridges, but even these cannot be that long in a single span. Suspended bridges were a step forward, with the improvement coming not from better beams, but from removing them altogether. Then, cable-stayed bridges solved some issues with older suspension designs and we can now have very large spans. Every time, scaling up is enabled by a change in strategy.

> Would a theoretical giant with a skeletal structure like a bridge or a house's frame be better off than one with a human-like skeleton of the same material?

Looking at things like elephants and dinosaurs, there certainly would be a way to have bones strong enough to support a 10m high human-like thing. Now, whether we would recognise that thing as human, I don’t know. Every time the problem was solved, there were significant compromises. I am not sure a giraffe’s vascular system could sustain a human’s brain. Elephants are not particularly agile. Whales cannot get out of the water.

Actual bone does incorporate some things we would consider "architectural" or "engineering" techniques to do more with less, not least of which they are composite materials and not mere rocks. They remodel the structure dynamically to respond to stresses as well.