For describing algorithms sticking with just the notion of graphs is convenient however, as it allows you to speak of edges, nodes, colours etc. keeping things easier to grasp.
Also, there are notions such as planar graphs which probably don't have a nice interpretation in the linear algebra point of view only.
It really depends on what is meant by graph theory vs. linear algebra here. If GT means we can fix an embedding to the space, then we have additional information such as number of crossings. However, it seems like a bit arbitrary restriction.
https://egraphs-good.github.io/
Addendum: many algorithms book describe algorithmic solutions to a multitude of graph problems without linear algebra at all, like connected components, network flow, shortest paths, etc, etc
P.s. what does being an industrial mathematician involve? Sounds like a role I would have dreamed about in undergrad.