 Hello, my name is Orel Kossron and I'm going to introduce the paper to our case-optimized hybrid homomorphic encryption featuring the Isabette swim saver. Hybrid homomorphic encryption, or HHT for short, is a tool used in fully homomorphic encryption to reduce the bandwidth usage by sending an homomorphic encryption of the key of a small symmetric scheme to the server and then using that scheme for every other transfer. If the expansion ratio of this red scheme is smaller than the gray, it's a win. But you then have to homomorphically get from the red box to the gray line server side an operation called transferring that can get some time to do. Several HHT schemes have emerged over the years, trying to optimize an aspect or another of the transferring, but all of them have measured their performances as standalone schemes. We, on the other hand, think an HHT scheme should always be considered with respect to a relevant use case, and with that in mind, we created a new scheme called Isabette as a successor of current state-of-the-art cryptosystem, Philippe. Elizabeth has been designed to be easy to evaluate on multiple threads and can store up to 4-bit messages in every safer text. Multi-bits data allow for an easier integration with a common FHT use case, neural network inference. We showed that, with some tweaks applied on the network prior to any encryption, it is possible to evaluate a small neural network with Elizabeth's safer text as input. These come at a cost on the inference time, which can be reduced with some bandwidth-costly optimization. However, using Isabette proved to be worthy from the very first inference, and the previously mentioned optimization are also rapidly compensated.