Engineering leader with 20+ years building ML and AI systems at scale. I bridge deep technical depth with organizational leadership — and operate fluently in both English and Japanese business environments.
Technical strategy, team structure, and delivery governance for AI-focused startups and scale-ups. Especially strong for companies bridging Japan and global markets — I work the technical and business sides in both languages.
Technical evaluation of AI systems and teams for VC and corporate investors. Architecture review, ML ops maturity, risk assessment, and capability validation for pre-investment and post-investment engagements.
Hands-on guidance for ISO/IEC 42001 implementation, audit trail design, and responsible AI frameworks for regulated industries — automotive, manufacturing, finance, and healthcare.
RAG systems, LLM ops infrastructure, workflow orchestration, and search/retrieval at scale. From whiteboard design through production deployment, with a focus on maintainability and observability.
Leading AI platform strategy, engineering transformation, and technical roadmap across automotive, manufacturing, and e-commerce sectors. Architecting CONFIDE, a multi-tenant RAG and workflow platform, and driving ISO/IEC 42001 AI management system compliance. Bilingual client engagement with Japanese manufacturers and enterprise stakeholders.
Led founding team that integrated ML into search. Delivered Mercari's first production use of LLMs and first customer-facing LLM feature. Designed and implemented ML Ops infrastructure and A/B testing for models. Spearheaded image search feature development.
Maintained high-volume ML systems with Kafka, Hadoop, Kotlin, ElasticSearch. Developed text intent platform for conversational automation. Built automated model training and deployment pipelines.
Introduced ML-based conversation automation, improving NLP capabilities. Architected microservices-based AI system with Django and Docker/Kubernetes. Built CI/CD pipelines for AI models and self-learning systems using human feedback.
Built NLP pipelines for multilingual document processing. Developed backend tools and translation support systems.
Available for remote fractional, advisory, and project-based engagements.