In this talk we will first briefly recall Grothendieck's etale fundamental group which was constructed in SGA1, then we will introduce Nori's generalization of the etale fundamental group for proper varieties via the language of essentially finite vector bundles. Next, we will introduce Nori's second idea on how to remove the properness assumption. But we will not completely follow Nori's second idea, instead, we will slightly generalize it and discuss some kind of Galois theory coming out of it.