OK, so look at what you have. Obviously [tex]f[/tex] is not exactly a tensor of type (1, 1), because it does not map a vector and a dual vector to a scalar; it maps a vector to another vector. The issue is further confused by the fact that [tex]f: V \to W[/tex] is a linear map from a space to a different space.
Nevertheless, there is a canonical isomorphism between the space of linear maps from [tex]V[/tex] to [tex]W[/tex], and the space of tensors that eat a vector in [tex]V[/tex] and a dual vector on [tex]W[/tex] and return a scalar. This isomorphism is what the question is asking you to find: give a recipe for converting a linear map [tex]f[/tex] to such a tensor, and show that this recipe is a one-to-one correspondence.
To see what to do, ask yourself: given [tex]f(v)[/tex] for some [tex]v \in V[/tex], which is a vector in [tex]W[/tex], what information would you need to turn it into a scalar?