August 2025 Summaries

Emiliia Kibenko's journey at DeepL illustrates the organic career growth possible in a supportive and dynamic work environment. Starting as a Software Engineer, she quickly advanced to a Senior Software Engineer role by embracing challenges, collaborating across teams, and continuously learning. DeepL's culture, which emphasizes a growth mindset, encourages initiative, open feedback, and learning from mistakes, providing a safe space for risk-taking and recognizing hard work. Emiliia highlights the satisfaction of developing and using the product, the balance between speed and quality, and the importance of curiosity and collaboration in career progression. Her story underscores how steady progress and a nurturing environment can make significant professional development attainable, inviting others to explore similar opportunities at DeepL.

DeepL has significantly enhanced its Large Language Models (LLMs) by transitioning from BF16 to FP8 precision, leveraging NVIDIA's H100 Tensor Core GPUs. The adoption of FP8, which uses fewer bits than BF16, has enabled DeepL to increase computational throughput and reduce memory demands without compromising the quality of training, despite FP8's narrower range and lower precision. This transition has accelerated the training process by 50% and allowed for the development of larger models with improved translation quality for European languages by 1.4 times and complex pairs like English and Japanese by 1.7 times, all while maintaining consistent latency for inference. By utilizing NVIDIA's Transformer Engine for mixed-precision training and TensorRT-LLM for inference, DeepL has effectively doubled the throughput capacity of its LLMs, enabling them to handle more requests and deliver optimal user experiences. This evolution signifies a substantial leap in scaling DeepL's Language AI capabilities, with future prospects of further advancements using FP4 tensor operations.