My attempt at a proof of this statement goes something like this: Let $(M,g)$ be a Riemannian $n$-manifold, then consider the vector bundle $\mathrm{T}M\oplus\mathrm{T}M$, then sectional curvature defines a function $K:\mathcal{U}\to\mathbb{R}$ which is defined on an open subset $\mathcal{U}\subset\mathrm{T}M\oplus\mathrm{T}M$ and is continuous, since it is smooth in coordinates. Note that we can define a $\operatorname{Gr}_2\mathbb{R}^n$-bundle on $M$ by considering a quotient topology on $\mathcal{U}$ in which we identify $(v_1,v_2)\in\mathrm{T}_xM\oplus\mathrm{T}_xM$ with $(w_1,w_2)\in\mathrm{T}_xM\oplus\mathrm{T}_xM$ precisely when they cut out the same plane in $\mathrm{T}_xM$. If we define the resulting quotient space to be $\mathcal{V}$, then we obtain a continuous map $K:\mathcal{V}\to\mathbb{R}$ and $\mathcal{V}$ is compact, so that sectional curvature is bounded.
However, I was talking to another student recently, and he claimed with great confidence that the sectional curvature on compact Riemannian manifolds need not be bounded, which makes me wonder if my (sketch of a) proof is incorrect.