Case Study: Positronic Robotics builds smarter cleaning robots with Nebius AI Cloud

How Positronic Robotics is creating a smart cleaning robot

Positronic Robotics, a startup creating AI-based robot control systems for cleaning, faced the challenge of training complex machine learning models to power its next-generation, autonomous cleaners. To build and refine its technology, the company turned to Nebius for its AI Cloud infrastructure, leveraging the vendor's Compute Cloud service with NVIDIA H100 GPUs for model training.

Nebius provided the scalable and flexible computational backbone for Positronic's development. Using Nebius' infrastructure, the startup runs 12-16 hour training cycles in containerized environments, which has been crucial for developing its advanced Action Chunking Transformer model. This Nebius-powered solution automates training processes and manages the storage of growing datasets, enabling Positronic to efficiently advance its robots' capabilities in performing precise cleaning tasks.

Positronic Robotics