Yeta

Meet Yeta

Yeta is a platform for adapting open-weight AI models and running them in production. Teams can bring their own data for supervised fine-tuning, preference tuning, or reinforcement learning, evaluate checkpoints, and deploy the resulting models for inference. Serving options include serverless endpoints, batch jobs, and dedicated deployments. Its runtime combines GPU virtualization with scheduling across different accelerator types, so workloads can run on a team's hardware, Yeta's managed cluster, or a combination of both. Developers can also use an OpenAI-compatible API and a web console to try models, manage API keys, and track usage. Yeta suits engineering teams that want model customization and serving connected through a shared infrastructure layer.