Ask me next year? I'm in the middle of the process, currently waiting for yet another round of denials from the local plans review people.
My office is purely off grid, too small to need any inspections, and works fine. I wouldn't go with my style of mounts for it, though. They're designed to swing, and... eh, don't build them that way. Not worth the hassle and weak points (the hinge to crossmember connections are showing some signs of stress, and I'm going to pull one of the arrays down, and redesign the other, while switching to a third style for the bulk of the panels).
If you're just doing a grid tied system, it's a whole lot easier than what I'm doing. I'm going with a ground mount system (16kW nameplate, give or take), but with east-west facing panels (longer solar day, better production on cloudy days), DC coupled to a 30kWh battery bank (for initial use - could increase later, but that covers my current desires for the battery bank), to a set of inverters that will let me run the house independent of the grid should I have the need to do so. Plus some transfer switches on the house and the like to let me bypass the system if needed. It's tricky, and I'm hoping the code guys give up in the next round or two and let me do it. I'm certain they've never seen anything like it before.
My power electronics will be in a shed, dedicated to this purpose (it gets me a NEC 690.12 exemption from panel level shutdown electronics in NEC 2017). The mounts will be large wooden A-frames (I tend to use wooden structures, painted, and it works fine out here) with Iron Ridge rails for the actual panel supports.
It sounds like you've done your research. I'm not familiar with the DC optimizers (I'm going with just plain strings of panels, though I have the option to toss in Tigo optimizers later), but I assume they'll meet your rapid shutdown requirements through some mechanism. If you're under NEC 2017, you'll need that if they run to the house for inverting.
Double check on the accessory structure code - if my mounts are >7' high, I need permits for them. Since I won't be doing that, I don't - but if you're going with taller mounts, you may need some local signoff. No idea where you live.
Check with the power company first. They're going to be the limit on system size, more than anyone else. I spent about a month and a half going around with my power company about my transformer size - the house transformer wasn't upgraded when we put our house in (should have been), so they were going to limit me to 15kW nameplate on a 15kVA transformer, even though it made no sense (I wouldn't have 15kW of inverter capacity). I now have a much nicer 25kVA transformer outside, which won't blow up were I to actually use my 200A service (the longevity of a 15kVA transformer supplying 48kVA is... less than long). Their self install checklist is helpful, but won't exempt you from meeting the workmanlike standards required in the NEC.
You'll also want to find that magical electrician who will sign off on homeowner work earlier rather than later, because they're rare beasts. Most simply won't do that. On the plus side, your wiring is fairly simple, so even having one do all the wiring work won't be hideously expensive.
I'm... roughly a college class worth of work into the project, and still hoping to get it done this year. Because I'm doing something more complex than a simple grid tied system, I've been deep in the NEC (have a printed copy), lots of calls back and forth to various suppliers and manufacturers to work out some corner cases, and as noted, am going around with the plans inspection people about some stuff. They don't tell you what they want, but are pretty clear I wasn't providing it. Worst case, it goes online next year and I lose 4% of the return.
For me, DIY is both about learning more (I'll be an expert on this type of system, and will be well documenting the build on my blog), and about having something I can't get anyone to build for a price I can afford. I'll have a $100k class off-grid capable system that can do load shifting and self consumption, for roughly want the local installers want for a grid tie system ($4/W was the rooftop system from some local installers, I'll be closer to $2-$2.50/W for the whole thing with the shed, batteries, etc). I could do a basic grid tie system out here for about $1.25/W, but it's just not that interesting to me. I'm not expecting my system to "pay off," but I'll have substantial grid independence, and the ability to optimize for whatever replaces net metering out here (it's going away soon, no idea what replaces it). And, with the east-west panels, a production curve that far better matches consumption, both locally and of the grid at large.
So, all that said, your best bet is going to be talking to the people who have done it locally. They'll be able to get you more useful local information. I can tell you the name of the guy that all solar roads in Idaho lead to, but it probably won't help you much.
