Isaac Asimov categorized three types of SF stories. There was Adventure Fiction/Space Opera in which the technology and science was secondary to the plot (which was often grandiose, epic or gonzo). There was gadget fiction in which the story hinged on the hero's attempts to gather resources and build a gadget to revolutionize ... something. Then there's social fiction in which the heroes must deal with the effects of technology on society. Put another way you can tell stories about your clone soldiers fighting the dark forces of the insurgency, you can tell a story about your struggle to build a clone-omatic chamber and acquire some Marilyn Monroe DNA, or you can tell a story about how your Marilyn Monroe clone refuses to take any more of /those/ kind of photos and is forming a group for clone rights.
SF roleplaying is mainly composed of the first and third types of stories. This should come as no surprise. Inventing new tech is pretty difficult and in most systems it is pointedly ignored or in some cases the subject of rules that make grappling rules look like the entrance exam for Pre-K.
Assuming you choose to allow inventions (and get through the rules) that leaves you with the fact that many (most) players will turn any tiny advantage into a monkey wrench aimed right at your campaign. The wrong kind of invention can wreck the feel of a campaign. Having your guys operate a grav apc when you're running an action hero game set on modern Earth would be a little jarring.
There are some ways around these concerns. First and foremost work out what you will allow with the egghead character's player in advance. Otherwise they might make that impossible roll and build that pocket antimatter battery.
Tech level advances are a relative thing. If the characters start out at TL 5 (World War Two era) then you might let them build a helicopter. The helicopter would give them an advantage; they could ferry commandoes around without using landing strips behind enemy lines. On the other hand helicopters are a well known piece of technology in modern Earth. You could look up the drawbacks (relatively slow and short range compared to fighters of the era) and requirements (lots of fuel and maintenance).
The same principle can be applied to any invention the players try to push on you. All technology has drawbacks. Prototypes have drawbacks you don't know about that could kill you.
A kind referee might allow a technology breakthrough just to level the field for his players (after getting tired of hosing them). A world facing alien invaders with superior technology might only have a chance to beat them by back-engineering their weapons and defenses. This is the whole point of the X-Com games. Kill the aliens and steal their stuff for the research teams.
Another option is to require a limited or hard to acquire resource to fuel the invention. Make the unobtainium the goal of one (or several missions). Or give the device an unpleasant side effect. Your starship is powered by psi energy. You have fantastic range and speed but the pilot powers it with his mental energy. If you push it too hard you might need a new pilot (or a new command crew if the pilot snaps.)
Finally new technology does not have to change the world (or galaxy or whatever) very much or very fast. If a ship has to clear 100 diameters of a planet before activating its jump drive new breakthroughs or fine tuning might allow your engineer to trigger a jump at 90 diameters (or 80, 70 etc). Not having to travel that ten diameters will save a trader time (and time is money to those guys). It can also allow you a quick retreat before pursuers know what happened. Neither of these will outright wreck a campaign. If other people see the players' ship keep jumping short of the limit they will probably start looking for ways to make their ships perform similarly and eventually everyone will be able to preform the trick and the technology war will begin again.