Scaling agentic AI: How UiPath built its high-performance GPU platform on AI Hypercomputer

UiPath built a high-performance GPU platform on Google Cloud's AI Hypercomputer to support its agentic AI initiatives, requiring massive computational power and infrastructure reliability. This platform allows UiPath to solve its 'spiky workload' problem and achieve predictable costs. The company re-architected its infrastructure to balance training and inference jobs, using A3 VM instances for training and G4 VM instances for inference. This partnership with Google Cloud enables UiPath to deliver high-precision intelligent document processing and autonomous agents that drive business outcomes. UiPath's engineering team utilizes heavy-duty math, including LLAMA model grounding and Qwen architecture, to power its robots and extract valuable data from paperwork.