Comparing developmental mechanisms across long evolutionary distances is challenging, especially when the level of description in the compared species is very disparate. This is the case for the gene regulatory network (GRN) that underlies the development of compound eyes in insects, where the very detailed level description of the network in the fruit fly Drosophila melanogaster contrasts with the knowledge of the network in other well studied insects, such as the flour beetle Tribolium or the desert locust Schistocerca. It is not even known to what extent the Drosophila eye GRN is typical of flies (diptera). Here we have introduced the marmalade fly Episyrphus balteatus (Syrphidae), separated from Drosophila approximately 90 MYrs, as a fly species to compare the mechanisms of compound eye development. The Episyrphus eye develops from an imaginal disc which is, in the overall, similar to that of Drosophila. By generating parallel data sets in Episyrphus and Drosophila to maximize comparability, our results show how the genes known to participate in the early specification of the Drosophila eye and the onset of its differentiation are likely participating in the eye GRN of Episyrphus, further expanding the set of conserved genes/nodes in this network. This gene set might be the basis for further explorations of the developmental mechanisms involved in compound eye development beyond flies. By combining RNAseq and ATACseq profiling with TF motifs we derive the first eye GRN in Episyrphus. When built using Episyrphus/Drosophila conserved nodes, the resulting GRN model exhibited abundant internal connectivity, suggesting that, as proposed in Drosophila, the GRN is knitted with many regulatory feedbacks. Using link conservation as a criterion for discovering new regulatory interactions, we find that the AML1/Runx transcription factor lozenge (lz) is a negative regulator of the retinal determination gene dachshund (dac) in Drosophila. However, although many of the predicted regulatory links are conserved, a large number of regulatory connections seem species-specific, even among conserved nodes. This suggests quite a significant regulatory rewiring during the 90Myr period separating Episyrphus and Drosophila.