This is definitely more than an art than a science so the right way to think about it is going to be different (sometimes dramatically so) for every company. My perspective is coming from working in Silicon Valley for the last decade. Some combination of these might be helpful to your situation.
  1. Is the person a cofounder?
    If so, then they should probably be getting comp in the same ballpark as the other cofounders, whatever that is (1/n of the equity, no or low salary until you have the funding to afford it, etc.) The tests I would use to tell if someone is a cofounder are: any of (a) would the company exist at all if she/he weren't part of it? (b) was he/she there at the beginning (surprisingly complicated question in practice!)?, (c) is the person such a value add that making them a cofounder helps everyone?
  2. The market rate for the position
    5 years ago, the rough market rate for a company making a first CTO/VP eng hire was 50bps (0.5% of the equity in the company) - 5%, depending on the company and the seniority and quality of the person. It's maybe a bit higher now because of the high demand for technical people and the general shift of power balance towards founders and early employees away from capital. For first technical hires not expected to be executives, the range is more like 50bps to 2%.
  3. Your plan for the company
    The girl or guy you bring in will be just one of a number of people you're bringing in as part of the team you're building to execute on your vision. You can't give everyone 10% of the company if you're hiring more than 10 people. Make a plan for how many employees you want to hire before you think you can reach a significant step change in valuation and then figure out how much each person gets as part of that plan.
  4. The market rate for the person
    On some level, the market for CTOs/VP eng is just that--a market. What other offers does he/she have (or would they have if they shopped themselves around)? You generally don't want to win someone over by paying them the most but there are a lot of cool things going on right now, and so you need to being offering something in the same ballpark as someone's alternatives if you want them to work with you. You can ask her/him for the value of their alternatives or ask your investors, or ask me.
  5. Expected Value Added Basis
    If you think someone will make your company X% more valuable, then in some sense, it's rational to give them anything up to the point where shareholder dilution is less than that. In practice, there's usually someone else who is close to as good who you would have to pass on hiring to bring this person on, so it's a little more complicated of a calculation. But you want to keep this way of thinking in mind so you don't miss out on non-traditional but super high value value adding opportunities.
  6. What Can the Company Afford?
    Most of these items focus on equity but salary is important too. If you are going to be paying someone a salary, figure out how that affects your runway. And remember that you might be setting a precedent around what other similarly skilled people will have to be paid if they work with you. Make sure you aren't setting yourself up to burn through all your cash before he/she can execute on what you need to prove to get more funding.
  7. The right thing for you is probably some combination of these things
    It's useful to triangulate on something that seems fair by finding some scheme that works under 2 or 3 (or more) of these frameworks. That ways it's also likely to be palatable to more of the people in your company and new employees and investors who might join you later.
  8. Other Details
    Everyone's equity should vest. Standard is 4 years with a 1 year cliff. For some very early people or highly compensated people it's occasionally 6 years. If people want more equity and have money, you can consider letting them invest as part of their compensation. Typically, salary and equity parts of comp. will have roughly similar expected value for technical people but that's not a hard rule. All of this can vary *a lot* depending on company, situation, etc.