Dasha is a platform of artificial conversational intelligence provided as a conversational-AI-as-a-service that allows companies and developers to incorporate realistic voice and text-based dialogue capabilities into apps and products.
With the integration of Dasha’s API, it is possible to build intelligent conversational apps for web, desktop, mobile, IoT and call center, explains the platform’s developer startup, which is currently in beta phase with the possibility to try
The voice is the native interface of human communication: then why do we continue to communicate with our technology using the fingertips? That is the question that has been asked and to which the founders of Dasha are trying to give an evolutionary answer.
A response that consists, as mentioned, in the possibility that the Dasha platform intends to offer to develop models of artificial conversational intelligence similar to human ones.
The versatility of this omnichannel solution means that companies can reuse the same dialogue through text and voice. In addition, its API-first approach allows companies to implement voice artificial intelligence models within call center environments, as well as integrate them into the web, mobile apps and IoT devices through Dasha’s SDK.
The graphic interface of Dasha Studio low code/no code, supports society, also allows those who are not a developer to create conversational apps of almost human quality.
To bring its innovations to the market quickly, Dasha performs all the training and artificial intelligence inference work for its conversation platform on Nvidia A100 Tensor Core and on previous generation GPU.
The A100 trains Dasha’s latest voice synthesis models in one day, seven times faster than previous generation GPUs. In one of his experiments, Dasha trained a Transformer 1.85 times faster using four A100s than eight V100 GPUs.