Technical Staff Member - Machine Learning Engineer / Scientist (Japanese Localization)

This Role Is Ideal For You If:You have a passion for developing LLM pipelines and agents for varied applications, and you thrive on troubleshooting edge cases where LLMs might struggle. As a native Japanese speaker, you aim to enhance LLM capabilities in the Japanese language. You are driven to tackle the complexities of adapting foundation models to different languages, cultures, and business workflows. Preferred Qualifications:In-depth knowledge of the Japanese model evaluation landscape and familiarity with Japanese pre-training data sources. Experience with modeling and inference tools such as Huggingface inference, vLLM, and cloud APIs. Your Responsibilities Will Include:Identifying, collecting, and curating diverse, high-quality datasets in Japanese text, audio, and multimodal formats. Creating synthetic methods to generate or augment Japanese training data as necessary. Ensuring that datasets fulfill enterprise-level quality, coverage, and compliance standards. Training and fine-tuning language and vision models to achieve top-tier performance for Japanese business applications. Modifying existing LFMs to cater specifically to Japanese language, culture, and business workflows. Establishing evaluation frameworks to assess model quality on Japanese datasets. Designing evaluation datasets and metrics tailored for Japanese enterprise applications. Conducting comprehensive error analysis and iteratively refining model performance. Ensuring robustness, fairness, and reliability in Japanese language outputs.