Why vector bundles are defined differently in algebraic geometry and topology?

Question

Over schemes, we can define

1. locally free sheaves;
2. vector bundles associated to a quasi-coherent sheaf (Stacks Project).

These notions seem to be dual to each other. It is easy to see on $\mathrm{Spec}(k)$ for a field, a locally free sheaf is a vector space $V$, and the vector bundle associated to the quasi-coherent sheaf $V$ is $V^*$.

Lemma 27.6.3 shows an anti-equivalence between the category of vector bundles and the category of quasi-coherent sheaves.

However topologically, vector bundles are defined first and its associated locally free sheaf is the sheaf of sections. This association is covariant rather than contravariant.

I wonder why algebraic geometers prefer this definition of vector bundles. Are there any conceptual or historical reasons behind this?