No, because the Ricci curvature is ($n-1$ times) the average of the sectional curvatures, so if they are all nonnegative it must be nonnegative.
* * *
More precisely, the result is false regardless of the compact Einstein setting.
We have, for every $p \in M$, for every unit vector $e_1 \in T_pM$, $$ \operatorname{Ric}(g)(e_1)=(n-1) \sum_{j=2}^n K(\operatorname{span}(e_1,e_j))=(n-1) \operatorname{Ave}_{\Pi \
i e_1}K(\Pi), $$ where $e_1, \dots, e_n$ is an orthonormal frame. Here $M$ is a $n$-dimensional Riemannian manifold.