I don't feel comfortable with using an external engine that I haven't written.
And yet you use a managed graphic API that you didn't write, which wraps a native graphic API that you didn't write, which calls drivers TYDW running on an operating system TYDW... dont forget .Net and the c# compiler TYDW.
So why are you not comfortable using an engine?
Jokes aside - as Sigil says Sunburn (or any engine) has no special access to anything and uses the same XNA framework and HLSL that you could write yourself.
My personal feel is that my time is worth money - lets say I'm worth $50 an hour. Could I write Sunburn in 5 hours? Nope - I'd probably take that long to get a solid shadow implementation alone and even then probably not as good as theirs. We're not talking about a few lines of code here... sunburn has been in active development by people with far more experience than you (dont take that personally, if you had to ask this question I'm guessing you are not an XNA or HLSL expert) for years and this is their full time job. Its way more than a few lines of extra code. Its thousands of lines and hundreds of hours of optimisation.
So for me an engine (sunburn or anything) is not about losing control - its about gaining time.
So if you are in this to finish the game then an engine is a great idea. Most folk don't realise how long a complete game takes and an engine can mean you can do more games in less time.
if you are in this to learn about graphics algorithms so you can maybe be an engine programmer yourself one day then go for it.
But bear in mind that engines dont solve every problem - e.g. sunburn has some 2d features but that is not its strengths. If it doesn't do what you need it to do then its worthless.