Then you definitely want a buck regulator, possibly even a buck followed by an LDO to clean up the output.
I've used TI's Webench tool to design a buck converter before, it's pretty easy to use, or you could design your own by hand if you're feeling adventurous.
That sounds like a lot of work for a one-off of a common device. Just go to ebay or your favorite far east merchant site and buy a "LM2596" buck converter module for <$1 shipped.
You could buy bobbins and spools of nichrome and wind your own resistors too, but most people just buy them.
You're welcome to do that, but I refuse to support the grey market/counterfeit industry that is so rampant on those sites, and I won't recommend it to others either. Not only are those devices notoriously overrated and unreliable, continuing to support that industry hurts legitimate manufacturers. Besides, he put efficiency at the top of the list, and I doubt he'd do much better with one of those modules than he would with a simple linear regulator (74%).
Have you ever used Webench? It's really not that much work, it does everything for you. Input your desired input voltage range, output voltage, and current capability, and it generates hundreds of designs that will meet it. You can sort those designs by BOM cost, size, efficiency, output ripple, etc. and then it gives you everything you need to know. Schematics, part numbers, BOM, graphs of everything you could want to know (efficiency vs output current and input voltage, thermal models, etc.), even the recommended schematic and board layout in a number of output formats (Altium, Eagle, etc.)
http://webench.ti.comFor an input voltage range of 9-11, output of 7.4, and 2A capacity, there's a nice LM3150-based design that will be over 97% efficient at .6A and up for a 1k BOM cost of $3.24, 500 mm^2 footprint, and 12 mV P-P ripple.