Interview with Kanishka Roychoudhury, GM of AI/ML at DigitalOcean
How does DigitalOcean cater to growing businesses?
“DigitalOcean is the developer cloud optimized for startups and growing digital businesses. We simplify cloud computing so businesses can spend more time on what matters. With our mission-critical infrastructure and fully managed offerings, DigitalOcean helps developers rapidly build, deploy and scale, whether creating a digital presence or building digital products.
By combining the power of simplicity, security, community and customer support, we ensure our customers can spend less time managing their infrastructure and more time building innovative applications that drive business growth”.
Why does the market need NVIDIA H100?
“The surge in interest for accelerated AI computing from businesses looking to elevate their capabilities in AI/ML has fuelled the demand for NVIDIA H100 GPUs. With their performance, these accelerators significantly reduce training iteration periods and model inference response times, making them ideal for running workloads such as large-scale AI model training as well as high-performance computing (HPC).
In response to this growing need, DigitalOcean now offers access to one of the market’s most powerful GPUs. This provides startups and growing digital businesses with state-of-the-art infrastructure crucial for developing the next generation of artificial intelligence/machine learning (AI/ML) applications. Through this offering and other recent enhancements, DigitalOcean is simplifying cloud computing and providing an affordable, end-to-end solution for developing, deploying, and scaling AI models”.
How does NVIDIA H100 differ from other solutions in the market?
“While many cloud providers are optimising their offerings to serve large enterprises, Paperspace by DigitalOcean offers AI/ML startups, technology companies, and growing digital businesses with reliable and flexible access to NVIDIA H100 GPUs. The Papespace H100 GPUs can be accessed on-demand or reserved for a specific term limit, allowing customers to select the most cost-effective consumption model for their business needs.
Whether AI/ML startups are looking to start small with a single H100 GPU, guaranteed long-term access to these next-gen GPUs, or large clusters connected by 3.2Tbps for their “bursty” spiky workloads, Paperspace is providing them with exactly the tools they need to build their AI/ML solutions quickly and reliably.
As we look ahead, DigitalOcean remains committed to providing solutions that offer superior performance and keep costs in check, lowering the barriers of entry for smaller businesses looking to leverage AI to improve profitability and efficiency”.
How will the H100 GPUs be used?
“Customers can access NVIDIA H100 GPUs as individual machines or as clusters, providing versatility for use cases ranging from multi-node model training to complex inferencing services. DigitalOcean has also made significant improvements to the underlying network infrastructure, now offering best-in-class interconnect speeds of 3.2TBps to enable optimal performance for users operating one or more of these eight-chip configurations”.
