Diamanti, creators of the first bare-metal hyperconverged platform for Kubernetes and containers, will share how large energy and service organizations are leveraging the Diamanti platform to validate their artificial intelligence (AI) and machine learning (ML) use cases at the ARC Industry Forum 2020 in Orlando.
AI and ML applications often leverage GPU processing for training models and they benefit from containers and Kubernetes. However, these processes are often complicated to adopt and run at scale. With the recent announcement of GPU support in the Diamanti AI/ML platform, enterprises have an easier on-ramp to managing large-scale containerized workloads under Kubernetes.
“We’re pleased to share the early customer traction we are seeing on our newest solutions in a wide range of industries including energy, services and more,” said Tom Barton, CEO of Diamanti. “These customers are validating state-of-the-art technologies internally while also benefiting from the reduced physical footprint and cost-savings that come with the Diamanti AI/ML platform.”
The new solution, announced in late 2019, is in early access today and fully supports Nvidia’s NVLink technology for higher performing workloads, as well as Kubeflow, an open source machine learning framework for Kubernetes that provides highly available Jupyter notebooks and ML pipelines. Combined with Diamanti’s Kubernetes control plane, this allows customers to deliver highly scalable environments for performance-intensive AI/ML workloads, accelerating model development and training.
Diamanti is seeing early customer traction in the energy and services sector with the early access release of this product, helping these organizations to deliver new AI/ML-enabled business improvements:
- A major energy company turned to Diamanti for a new workload leveraging AI/ML for optical character recognition (OCR) to scan invoices. The customer needed to scan more than 15,000 invoices a day. The legacy infrastructure could not keep up with the demand and eventually accrued a backlog of more than 200,000 invoices. Deploying the Diamanti solution with GPU support eliminated that backlog within hours.
- A major online travel company wanted to double its conversion rate on website visitors looking at offers and booking travel. They are using AI modeling to determine the best approach, but unsuccessfully ran proof of concept trials with several major vendors before turning to Diamanti. With Diamanti, they are meeting their new performance requirements without any dependencies on proprietary software. They can also run both their GPU-dependent applications and standard containerized applications in the same environment, minimizing the complexity of managing multiple stacks.