Germany-based AI-powered language translation company DeepL is keen to strengthen its presence in the Middle East, and says it is working to improve its Arabic translation services while tapping into potential customers in the region. said the company's CEO.
The company's AI neural translation technology has set the pace in many areas of the language translation field in recent years, said CEO Jarosław “Jarek” Kutylowski, 41. The National He said this in an interview in Dubai.
The technology translation unicorn's efforts to increase its presence in the region come amid a global race for a piece of the highly profitable language translation sector, with companies as diverse as Google, Microsoft and Grammarly.
Kutilowsky said DeepL offers deep VC funding and hundreds of millions of users and is trying to stay ahead of its competitors.
DeepL Pro, its translation service for enterprise customers, is used by 20,000 companies, according to DeepL.
“We are starting to talk to our customers to see what their requirements are. We are keeping an eye on the UAE and will see what happens in the future,” he said.
Kutilowsky said that DeepL is in talks with various companies in the UAE, particularly in the real estate sector, to learn more about how it can help international customers.
He also strives to better understand the nuances of Arabic in the business world.
“One of the reasons I'm here is [in Dubai]I am interested in this language and hope to learn more during my time here,” he said.
Headquartered in Cologne, DeepL currently has offices in London, Amsterdam, Berlin, Tokyo and the US, with no immediate plans to open an office in the UAE, although that may change.
“It depends on the success of the market and the reception from customers and how much support potential customers need locally,” he said.
Founded in 2017, the company reached a valuation of $1 billion last year and launched Arabic translation capabilities in December of the same year.
Despite advances in machine learning, Arabic has proven difficult for various online-based translation services due to the complexity and diversity of its dialects.
Kutylowski said Arabic is the first right-to-left language translation implementation launched by DeepL.
“The grammar is a little different and some things in AI translation training need to be done a little differently,” he said.
“The advantage of AI is that you don't have to teach it grammar rules and it already understands them, but there are some preparatory steps that require additional work to complete it,” he said, adding that it also understands Japanese. I pointed out that it is the same. It took even more time to perfect it.
Kutilovskiy said implementing Arabic translation is also a matter of priorities and staffing, and that DeepL is trying to put that front and center.
“It's work,” he said. “From an engineering perspective, it's not that complicated, but it just needs to be done.”
He said DeepL's workforce is approaching about 1,000 employees.
“This is a complex topic, so 60% of it is research and development.” [translation] We also have a large platform for designing. We're in the top 100 domains in the world in terms of traffic, and we want to keep it running consistently. ” he said.
“There's also research on neural networks,” he said, noting that DeepL has about 50 people working on the mathematical and technical aspects of such networks.
Kutylowski also said the company employs editors and others who provide input in the 32 languages that DeepL supports.
According to the company's website, 20,000 companies currently use DeepL for customer and internal communications.
Kutylowski listed the Japanese news agency Nihon Keizai Shimbun as one of DeepL's customers.
“They translate some of the articles written in Japanese into English and Chinese and publish them,” he said.
DeepL translation is made possible through artificial neural networks, a key component that enables artificial intelligence. This means that these networks used in most modern translation apps and services are trained in a similar way to the human brain.
According to Kutylowski, DeepL was one of the first to use this approach, and while other translation services have also used neural machine translation in recent years, DeepL approaches this by focusing on training data and methodology. has been improved.
“If you can design the best neural network architecture, train those networks properly, and find the best training data, you can achieve the best quality, and that's what we've been working on for six years. ” He said.
In addition to free basic translation services, DeepL offers various iterations of translation services, apps, services, and APIs through tiered pricing.
The company also recently announced what it calls an AI-powered writing companion called DeepL Write. This puts it toe-to-toe with Grammarly, a similar product widely considered to be a dominant player in the writing assistance space.
Kutylowski said most of DeepL's customers use desktop and laptop PCs and DeepL's browser extensions, but mobile apps are also available.
“This is still the status quo in businesses and enterprises,” he said, noting that users prefer to use DeepL's services on PCs rather than mobile devices.
“Most of your work during the day is done on your desktop when using advanced applications.”
Kutilowski did not provide details on revenue, but called DeepL's business model and financial outlook “good” and said the company's web traffic consistently ranks among the top 100 largest web domains in the world. Ta.
When it comes to providing online translation services, DeepL is not alone. Competitors such as Google Translate and Microsoft Bing's translation services also offer Arabic translation services, although not in the personalized client model offered by DeepL.
Google and Microsoft also have deep pockets to improve search over the long term and an installed user base that could potentially be a big advantage.
The translation market shows no signs of slowing down. According to Straits Research, by 2031 the market size could reach $48 billion.
DeepL continues to attract investors and customers while expanding its language portfolio.
“We're very good at focusing on what we're doing,” Kutilowsky said, focusing on DeepL's neural network research and staffing language and engineering experts. He mentioned this.
Jarek also made sure to point out that DeepL was his first company and he has no plans to move on to another company anytime soon.
“It's hard enough to track all this growth, so we're not going to get sidetracked,” he said.
Updated: May 9, 2024, 3:00 AM
