Some general principles I follow for almost any race. There's definitely more aggressive playing styles possible but this tends to work for me (difficulty ranging from painful to crippling, depending on race and galaxy size):
1) Research cheap technologies (let's say maximum 5-10 weeks at 100% production and 33% research) that increase social production to develop your first few planets sooner, e.g. xeno engineering, planetary improvements
2) Research cheap technologies that increase military production to build colony ships faster, typically the first 1-2 in the red branch
3) Research cheap technologies that increase research, e.g. xeno research
4) By that time I either have sufficient research capacity to repeat 1-3 for another round or I'll go after some speed techs to improve my colonization speed, e.g. ion drive or impulse drive
5) Next it's down the yellow branch until 'trade' so I can get an econ capital up, followed by xeno industrial theory to get a manufacturing capital
6) To be ahead of most if not all AI's in the early weapon rush, down the weapon branch to say laser, then medium scale building. If you have access to total majesty, this would be a good time to research until there and build a spin control center
7) When I have a handful of labs, advanced computing to build a tech capital
At some point I usually need to improve my cash flow so I'll research anything quick that gives a boost to economy, morale or population growth. Sensors are also a good choice, then build a few more surveyors who go after the 1000 bc anomalies.
After that it's usually time to research better weapons and some miniaturization, potentially large hulls. If those techs would take too long, improve your labs first.
Of course, you can get your hands on many of these technologies by trading with the AIs, by either offering them some technologies of your own, cash (though not too much early on in the game since you'll need that cash to stay afloat until your colonies become more profitable) or influence points (never more than 999 at a time, due to an error the AI doesn't value 1000 or more influence points at a time)