#None #NVIDIA #DeepL #Language AI

DeepL's Revolutionary Leap in Language AI

In an exciting development, DeepL, a prominent player in the Language AI industry, has announced the deployment of an NVIDIA DGX SuperPOD equipped with DGX GB200 systems in Europe. This powerful computing platform significantly enhances DeepL’s ability to translate the entirety of the Internet in just 18.5 days, compared to the previous time of 194 days. This remarkable upgrade not only boosts translation speed but also paves the way for advanced research and innovative feature development.

Enhanced Translation Capabilities

With the new NVIDIA DGX SuperPOD, DeepL is now capable of processing translations with impressive accuracy and speed. For instance, the time required to translate the Oxford English Dictionary has drastically reduced from 39 seconds to an astonishing 2 seconds. Furthermore, the translation speed for Marcel Proust's lengthy novel In Search of Lost Time has improved from 0.95 seconds to just 0.09 seconds. In total, DeepL can now produce 30 times more text than it could previously, marking a significant milestone in the realm of Language AI.

Research and Development Advancements

The introduction of this state-of-the-art SuperPOD is more than just a technical upgrade; it represents a boost in DeepL's research capabilities, facilitating the development of powerful generative features that are set to expand its product offerings substantially. The advanced infrastructure provided by the DGX systems will allow DeepL to approach model training in innovative ways, ultimately leading to a more interactive user experience.

In the short term, users can anticipate enhanced translation quality and speed, as well as the introduction of new generative AI functionalities, improving interactivity during the translation process. Looking ahead, DeepL aims to integrate multimodal models into its services, allowing for a tailored experience that meets the unique needs of each user.

Technical Specifications

The SuperPOD is designed for effective scalability, equipped with liquid-cooled racks that house tens of thousands of GPUs. This allows DeepL to operate high-quality AI models that are essential for next-generation generative applications. As a successor to the previous DeepL Mercury supercomputer, the NVIDIA DGX SuperPOD marks a substantial step forward in computational power.

CEO Jarek Kutylowski stated, "At DeepL, we constantly invest in research and development to provide superior solutions compared to our competitors. This new supercomputer cluster will further solidify our position as a market leader in Language AI. By equipping our research infrastructure with the latest technology, we not only enhance our current offerings but also explore new exciting products. The pace of AI innovation has never been higher, and integrating this new knowledge into our tech stack is crucial for our continued growth."

The Need for Advanced AI Infrastructure in Europe

As the demand for robust AI clusters grows, experts like Charlie Boyle, Vice President of DGX systems at NVIDIA, emphasize the necessity for Europe to maintain its competitive edge. By leveraging cutting-edge AI infrastructures, companies like DeepL are well-equipped to accelerate breakthroughs in Language AI and deliver enhanced user experiences throughout Europe and beyond.

About DeepL

Founded in 2017, DeepL seeks to break language barriers globally. Currently, over 200,000 businesses and governments, along with millions of individuals across 228 countries, utilize DeepL's Language AI platform for natural sounding translations and multilingual communications. The platform has earned a reputation for its advanced data security and AI tools, transforming business communications and enhancing productivity.

DeepL continues to grow, bolstered by support from renowned investors, and aims to innovate continuously in the competitive landscape of AI-driven language solutions. For more information, visit deepl.com.