Carnegie Mellon University
September 24, 2012

Press Release: Carnegie Mellon Spinoff Safaba Automates Translation for Global Companies

Customized Machine Translation Boosts Global Corporate Communications

Contact: Byron Spice / 412-268-9068 / bspice@cs.cmu.edu

PITTSBURGH—Safaba Translation Solutions, LLC, a spinoff of Carnegie Mellon University’s Language Technologies Institute, has launched a new service that lets companies directly tap the power of automated translation to enhance global communications and marketing.

Safaba customizes state-of-the-art machine translation to conform to each client company’s branding and the vocabulary of its business domain. These Safaba-engineered systems can significantly reduce the time necessary to produce translated text, while also delivering translations with far higher quality than possible with generic online translation programs.

Alon Lavie, research professor in CMU’s Language Technologies Institute and the president and CEO of Safaba, said statistical machine translation technologies, once optimized for a client by Safaba, can easily double translation output.

Greenlighting Logo“Human translators typically produce about 2,500 words per day,” Lavie said. “With Safaba’s machine translation, we can increase that velocity to 5,000 or 6,000 words per day. This allows Safaba’s clients to migrate a more complete brand to global markets and deliver the customer experience they want.”

Though machine translation technology promises significant increases in translation productivity, fine-tuning the solutions takes knowledge beyond the capability of most companies and even most translation firms, noted Lavie, who earned his Ph.D. in computer science at Carnegie Mellon.

“Developing machine translation systems isn’t for the fainthearted, but our people have intimate knowledge of this technology, both as users and developers,” he said. “We know how to tear apart these open-source systems, add customized components and re-assemble them in ways that optimize performance to meet a company’s specific needs.”

Safaba operates on a software-as-a-service model, hosting and maintaining software on its own servers that clients can access remotely. Its systems already provide translation for 22 languages, with another five languages set for deployment soon.

The demand for translation services has grown as markets have become global and the volume of text to be translated forces even large companies to rely on freelancers and outside vendors. Safaba offers these companies the option to bring more of their translation work in-house and to exercise greater control over content creation, Lavie said.

In the case of marketing materials and other documents requiring human-level quality, the translations by the Safaba systems provide a means of increasing the speed and volume of translations, which can then be edited by human translators. For customer support documentation, which often is too costly to be translated manually, the translations provided by Safaba may be of sufficient quality to be distributed in their raw form.

In addition to Lavie, the management of the Pittsburgh company includes Vice President of Engineering Robert Olszewski, who earned his Ph.D. in computer science at Carnegie Mellon, and Vice President of Business Development Udi Hershkovich, who has nearly 20 years of experience in the software industry.

Safaba has received support through the School of Computer Science’s Project Olympus and Carnegie Mellon’s Center for Technology Transfer and Enterprise Creation. They are two of six campus incubator groups, collectively known as Greenlighting Startups, which support CMU’s culture of entrepreneurship. An engine for accelerating innovation and job creation, Greenlighting Startups builds upon the university's impressive record of turning campus innovations into new businesses by supporting award-winning professors and world-class students in transforming their research into thriving commercial enterprises.

The company also has been the recipient of funding through the National Science Foundation’s Small Business Innovation Research Program, and from Idea Foundry and Innovation Works of Southwestern Pennsylvania.

###