Case Study: NVIDIA improves multilingual content quality and cuts costs with LILT

NVIDIA Improves Multilingual Content Quality 30% and Cuts Costs 32% With LILT

NVIDIA, a global leader in accelerated computing and artificial intelligence, faced challenges with its manual and fragmented multilingual content workflows. The company was using four different translation platforms, leading to inconsistent processes, redundant spending, and an inability to integrate its own AI models. NVIDIA selected Lilt to help integrate its Nemotron speech model and build a more efficient, AI-first translation system.

Lilt implemented a solution that integrated NVIDIA's Nemotron speech to generate initial translations, which were then refined by Lilt's adaptive AI and expert human reviewers. This partnership improved the quality of NVIDIA's translation model by 30%, reduced costs by 32%, and doubled the volume of localized content. The consolidated and automated workflow also allowed NVIDIA to translate over 12 million words in a single year.

NVIDIA