Responsibilities
- Own the design, development, and operation of AI-native features centered on RAG, LLMs, and AI agents, going beyond model implementation to deeply integrate with UI/UX and business workflows to deliver truly usable AI experiences
- Lead end-to-end decision-making across model selection, inference architecture, RAG design, and agent design based on specific use cases, balancing both quality and speed
- Design and implement robust operational frameworks—including evaluation, monitoring, logging, and cost optimization—to continuously improve and maintain a reliable and trustworthy AI experience
- Design and implement highly reliable agent-based LLM workflows for production environments
- Build execution infrastructure for AI agents by integrating with external services and internal APIs
- Evaluate and select appropriate external models, frameworks, and services based on use case fit
- Translate user needs into concrete requirements through stakeholder discussions, and collaborate cross-functionally to design optimal agent-based solutions
- Establish testing frameworks and monitoring systems to evaluate agent performance, and visualize key metrics
- Define development processes and quality standards through code reviews and comprehensive documentation
- Identify and address technical challenges to ensure the accuracy, reliability, performance, and scalability of AI systems, driving continuous improvement
Requirements
- 5+ years of development experience in at least one of the following: Python, TypeScript, or Go
- Hands-on experience building applications using LLM APIs (e.g., OpenAI, Anthropic, Google Gemini, Mistral)
- Practical experience developing AI agents (e.g., LangChain, CrewAI, OpenAI Agents SDK, Google ADK, MCP) and working with context engineering
- Experience designing and implementing RAG systems, including embeddings, vector databases, and retrieval algorithms
- Experience improving generation quality through prompt optimization and the design of evaluation metrics
- Solid foundation in natural language processing (NLP) and experience building data pipelines
- Japanese proficiency equivalent to JLPT N1
Nice to haves
While not specifically required, tell us if you have any of the following.
- Experience building and operating search infrastructure using Elasticsearch, OpenSearch, Vespa, or similar technologies
- Experience designing systems that integrate knowledge graphs or structured data with RAG
- Experience designing and operating systems in cloud environments such as AWS, Google Cloud, or Azure
- Experience preprocessing and normalizing unstructured text data, such as customer emails, FAQs, and product documentation
Compensation
¥7,700,000 ~ ¥15,000,000 annually.
