Case Study: TUNiB scales its emotional chatbot service with FriendliAI Dedicated Endpoints

How TUNiB easily managed & scales their emotional chatbot service

TUNiB, a generative AI startup, faced challenges in managing and scaling the infrastructure for the large language models powering its emotional chatbot service, Dearmate. The engineering effort and operational costs for LLM maintenance were substantial, requiring specialized skills in deep learning and distributed systems. To address this, they turned to the vendor FriendliAI and its Friendli Dedicated Endpoints solution.

FriendliAI provided a solution with dedicated GPU instances, ensuring consistent performance and handling resource allocation, scaling, and monitoring. This allowed TUNiB to deploy custom models and focus on its core application. The results included a convenient, reliable, and cost-efficient service, with TUNiB able to onboard its chatbot service in less than 20 minutes and achieve uninterrupted operation.

TUNiB