I know how to prove if $A,B$ invertible then $AB$ is invertible and $(AB)^{-1}=B^{-1}A^{-1}.$ But I'm not sure about how to prove the converse.
My attempt:
If $AB$ is invertible, then $B$ must be one-one and $A$ must be onto. Since $\mathcal L(V,W)\sim M_{n\times n}(F)$, $L_A:V\to W$, $L_B:V\to W$ and $\dim(V)=\dim(W)$ and they're finite dimension and linear, so one-one means onto and onto means one-one, so $A,B$ are invertible.
Is this correct and/or how to improve it?
Your proof is fine!
If one can prove $B$ is one-one and $A$ is onto, then the result follows from this fact:(as you mentioned)
Suppose $AB$ is invertible, then $\exists C \in L(\Bbb{F}^n)$ so that $$(AB)C=C(AB)=I_n$$
Let $v\in F^n$ so that $Bv=0$.
$v=Iv=(CAB)v=CA(Bv)=CA(0)=0$,concluding $\text{null}\; B=\{0\}$ and so $B$ is invertible
Likewise, let $v \in F^n$ be arbitrary.
$v=Iv=(ABC)v=(AB)Cv=A(BCv)$, concluding $A$ is onto