Wouldn't quite classify myself as a deity player yet, but I can tell you some thoughts based on my game.
I went with Mehmed for cheap workers and granaries, didn't plan on bulbing anything. Org probably came in handy as well as I didn't go bankrupt. In the final save it is saving me 25gpt, which isn't much, but it makes the difference between running a deficit or a surplus at 0% slider. Though the main key to not going bankrupt is to keep that conquest gold rolling in. Never slow down unit production, keep reinforcing your army so that you always can move onto the next target as soon as possible.
Not sure about best settings. I went toroidal aiming at micro pangaea. My map was 360 tiles, which I later learned isn't one. The small micro pangaeas can have as few as 240 tiles. However, when I later tried a few games with maps like that (for a different slot), I noticed the AI would expand to 2-3 cities, run out of space and stuff to build, then spam tons of units. Doesn't necessarily make it easier. I don't know about tile counts for flat or cylindrical maps, but if it doesn't go well over 400, it might be worth it to go with one of those.
All I know about worker stealing is that it gets much harder on deity. It's harder to get peace and harder to prevent the AI from hooking up metals. I stole one from Gandhi and had to give him writing for peace when I saw him hook up copper with no possibility to stop him. The nice thing with Gandhi is that he was pleased with me almost immediately after taking peace.
Can't say much about optimal tech order, depends on map and how you are planning to win this, I suppose. I popped pottery from a hut and went mining-writing-alpha, then backfilled through trades. Oracle construction would sure be nice though, if you have elephants. Or maybe oracle currency.
I'd recommend elepult as the preferred mode of war, if you are not very experienced with early deity warfare. It's ridiculously strong before feud and engineering. I always run into economical problems if I try HA-rushing on this level. If you have a neighbor without metals and find iron yourself, some early swords could be an interesting option to grab some more cities and forests for elephant production. You can usually trade for IW early, so it doesn't delay key economical techs like HBR does.
If you can keep the AI busy fighting each other, it helps a lot. I got lucky as they took care of that part themselves. Freddy even DoWed Gandhi while they were sharing religion. That came at the best possible time, as I saw him move about a dozen random metal units into Gandhi's territory right before I attacked him. His stack never came back to defend his cities.
Then it just comes down to building lots and lots of units. In my game I built 49 elephants and 31 pults. Maybe a bit excessive as 30 elephants were alive still at the end and almost 20 elephants didn't fight at all in the final war against Gandhi. Really should have attacked him earlier, but I was so afraid of his castles, pikes and longbows that I waited until all of my army was in position before declaring.
But I had elephants. Missed oracle construction by 1 turn... But man those elephants were great once I finally got there. 
.
