ロータスジャパン株式会社

English voice data annotation service for segmenting English sentences into words based on waveforms. Assembled a team of 40 members to annotate 80,000 sentences with 95% accuracy through quality assurance processes.

Created a modular RAG pipeline architecture with automated testing for multiple RAG configurations. Developed evaluation metrics dashboard and integrated cloud services for data storage with auto-deployment to chatbot interface.

AI-powered solution to standardize and localize professional profiles across markets. Implemented LLM-powered content generation pipeline with translation capabilities (92% accuracy), intelligent template manager with cultural adaptation, and multi-format document generation system. Supports 10 concurrent resumes with <1s API response.

Custom RAG chatbot system deployed on Mistral-7B using healthcare provider's database. Incorporates RAG development, LLM model integration, and container deployment to API services for virtual customer assistant providing quick medical advice.