Mirendil Leverages Google Cloud's AI Hypercomputer for Frontier Research

Mirendil, a frontier AI research company, has expanded its use of Google Cloud's AI Hypercomputer to support the pre-training and post-training of advanced artificial intelligence models. The company will utilize a combination of Google Tensor Processing Units (TPUs) and NVIDIA accelerated computing infrastructure running on Google Cloud.

The AI Hypercomputer provides purpose-built infrastructure spanning accelerators, software, networking, and storage for model training, inference, and advanced research. Mirendil plans to use this infrastructure for complex model-training workflows, including initial pre-training, post-training, and large-scale reinforcement learning.

According to Behnam Neyshabur, Co-Founder and CEO of Mirendil, the goal is to accelerate the research process itself by developing AI systems that can help scientists and engineers design experiments, evaluate results, and iterate more efficiently. The company aims to make frontier AI research capabilities available to a larger group of scientists and engineers.

Google Cloud worked with Mirendil on the design and deployment of the combined infrastructure across computing, storage, networking, and control planes. The companies also collaborated on a system that uses managed training clusters running through the Gemini Enterprise Agent Platform, which is intended to streamline provisioning and management of Mirendil's TPU and GPU environments.

The deployment gives Mirendil access to flexible computing resources as it develops AI systems capable of improving and automating portions of the research cycle. Mirendil is already operating a cluster powered by Google Cloud TPU v5P chips, with NVIDIA accelerated computing systems expected to come online soon.