What tactics regarding semi-complex project structure should I take?

I did google, but I feel that all info found is either non-understandable (too generic for my skill level) or not applicable to me at all.

BACKGROUND
Most of my videos so far were simple music-video-clips (video with music in background), lasting approx. 5 min, and 1 being a 10 min vlog-style interview (talking head, then 20-30 sec music-video sequence, then back to taking head, and repeat this pattern several times). However, I need to take my edits to a next level.

NESTING AS I KNOW IT
I've read about nesting, and that it's best to use non-proxy clips in nested projects, then such sub-projects are nested into the main project file, this time round as a proxy clip for smoother preview during editing.

NESTED SUB-PROJECTS ISSUE 1 - TIME CONSUMING ITERATIONS
My concern is that if, during editing of the master project, I realise that one of sub-projects need re-adjusting (likely several iterated attempts in order match, say, "cut on motion"), the entire workflow might stretch enormously in time.

NESTED SUB-PROJECTS ISSUE 2 - MOVING FEW CLIPS OR ONE WITH CHANGED SPEED OFTEN SCREWS UP CUT/TRIM
And there is another twist: I've noticed that Kdenlive can not cope when there are too many clips on timeline, and even screws-up timeline entirely if few clips are selected and moved together. My educated guess is that Kdenlive is not best in managing the tags if the TXT project file is too large - on my computer the treshold size of a project file is approx. 1 MB. Worse still, if any of moved video clips has speed different than original 100% (typically for Fast-Slow-Fast effect; and why isn't such effect OOTB in Kdenlive anyway?!), then the trimming of such video clip goes out of the window. I could pre-render those clips with speed change, but...

ISSUE WITH PRE-RENDERED SUB-PROJECTS
Sub-projects rendered and then imported (not nested), will certainly take less time on the final render, and Kdenlive won't be able to screw them up, but there will be a quality loss (otherwise those sub-renders will eat my disk space) and there is zero flexibility for any re-adjustments.

MY CLEVER IDEA
I came up with a tactics that the interviewed talking head will be in a nested sub-project with several gaps of black-background, all B-roll music-video-clips as sub-projects. Then all sub-projects imported into a master project file, the interview nested sub-project split into several parts, and gaps filled with music-video nested sub-projects. If I need an extra B-roll whilst the talking head is still talking, I'd just add relevant video clips directly into the master project timeline.

WHAT SHOULD I DO?
So, the simple question is: what is the best tactics/balance between nesting, pre-rendering, and not nesting? Won't Kdenlive mess up trimming if a nested sub-project is split into several parts inside the master project, and those parts separated with all other nested sub-projects?

Thank you very much in advance.

Anybody, please?

I cannot answer your question as I am very much a newbie. In fact, I have a much simpler related question that lead me to yours. Maybe someone can help us both.

I make training videos for EMS and am finding that Kdenlive is way better than any other I have found. Now that I can do the simple things, I find myself wanting to do more (good I think) but that adds complexity. I THINK I want to nest one project inside another as a way to keep the number of clips, effects, etc., from growing too complex. Simpler seems better because I don't get lost in long time lines and, frankly, I have less chance to mess up portion that is already working well. I guess I could just have two projects and render each to video then join those two long clips into one and then render the result. But back when I used to do audio-only editing (analog of course), that was a bad idea because each generation added its own degradation.

What's wisdom can someone help me with? Thanks

20.06
Total length: about 30 mins
Clip count: maybe 40 - 50
Oh, yeah, still using Windows 10 (but for many reasons probably going to upgrade machine and run a Linux distro

/s/ John