I'd go work as a software engineer for 3-5 years, work on the LSAT until you're at least at a 168, and then apply to law school.
Why a software engineer over USPTO?
If you look at most of the patent
attorney job postings that are aimed at CE/EE majors, they seem to fall into two groups: 1) those that like USPTO experience, and 2) those that like actual engineering experience. It appears that larger firms like actual engineering experience over USPTO experience and that boutiques and other firms/positions like USPTO experience over engineering experience. I'm not exactly sure why, but that's what it looks like from job postings and from those I've spoken with in the field.
As a general rule, actual engineering experience will help you with the more advanced patents and will open unique doors. For example, there is a firm in DC that does nanotech patents. I forget the guy's name, but he was the first on the scene for that field in patent law (Phd from Hopkins and JD from UVA). He's said that he looks for people with actual practice in the field because his firm's work is so specific that it requires a high degree of familiarity with the material. Other bigger firms -- like Kirkpatrick -- like both types of experience, but tend to utilize USPTO experience more heavily (that's from my friends who work there). They say that they utilize their engineering degree/experience about 20% of the time, and spend the other 80% using wikipedia to figure something out because their subject matter can be rather broad. I've also heard this from professors who specialize in patent law.
I personally believe that a good route through patent law is to work at a large firm for a few years, make excellent money, pay off your debts, and develop a nest egg. Then you can transfer into a boutique or an in-house position at a tech firm. These provide a good balance of hours and a healthy salary. For this path, from what I understand, it is best to have work experience as an engineer. If you'd prefer to go straight to a boutique or straight in-house at a tech firm (both are more rare than the former), then USPTO work seems to be a better choice; I'm guessing because it familiarizes you with the process, rather than the substance, which appears to be more important at smaller firms.
Why a 168?
It looks like the number that will get you to be considered at a series of solid schools, and might even get you a scholarship at some of the T50. Any higher would be great, but from speaking with my friend who used to teach LSAT prep, a 168 is usually a peaking score for most folks, so I'm just trying to be realistic. I would personally cut off at anything below a 165. At that point you likely won't get into a T50 at all, especially with any money. Of course that's my personal cut-off. Pretty much anything in the T100 could still be rationale with engineering WE and a substantial scholarship. It would be irrational for most people, but patent law is a different animal. I recently read that for each regular
attorney job opening, there are 30 patent attorney offerings. I never checked that information's source, but it appears to be feasible on its face (look at OCIs, job banks, etc. and you should see a similar correlation). You also have a fall back career that could pay off your debt without substantial difficulty.
Of course I'm simplifying this cost-benefit analysis, but there are too many variables to be certain either way. But to summarize my opinion, I'd consider two options: 1) a T50 school with no more than 200K in total debt by the time of graduation, and 2) a T100 school with no more than 100K in debt.
Again, most people will balk at these numbers, but patent law is a different beast. You can realistically come out of a T50 school at median with a CE/EE degree and WE and secure a patent
attorney job starting at 160K in the first year and up to 300K+ by the fifth or sixth year. And you can do that in a small town that has a low COL. It's not easy, but it's definitely feasible. I know because I have several friends who are doing it. With a salary like that, you can pay off that 200K in as fast as 3 years, but more realistically somewhere between 5 and 10 years (dependent on your lifestyle).
The same type of analysis goes for the T100 and 100K cutoff. And if shit hits the fan and you fuck up law school entirely, you can still become a CE or EE with several years work experience, which should allow you to pull in roughly 100K, dependent on the market. That's definitely enough to pay off 100K in 5-10 years also.
Either way, best of luck with your choices. I've had to do far too much research into this stuff because I'm doing it all backwards (picking up my entire physics degree during my rising 3L summer and over my 3L year so I can sit for the patent bar once I graduate). So if you have any questions, feel free to PM me. If I don't know the answer (more than likely), I probably know someone who does.
edit: I meant Kirkland.