Research Engineer

JAPAN AI (a Geniee group company) is hiring a Research Engineer to work on its enterprise AI agent platform, JAPAN AI STUDIO. Responsibilities include conceiving, developing, and comparing agent harnesses (memory, context compression, inter-agent communication), designing quantitative benchmarks for large-scale agentic tasks, supporting automated model/prompt evaluation, optimizing inference latency and cost (quantization, distillation, caching), building training data mixes, and transferring research outcomes to the product engineering team. The focus is applying research directly to a production environment serving roughly 200 companies rather than publishing papers, though paper/blog publication and academic collaboration are supported. Candidates should hold a Master's or PhD in CS, ML, AI, or a related field, have hands-on experience building agentic LLM systems, strong production-quality Python skills, and experience with large-scale model training/inference using PyTorch or JAX. At least one of Japanese (fluent) or English (business level) is required. Salary is ¥12,000,000-¥20,000,000/year (includes 45h fixed overtime), with stock options and twice-yearly bonus reviews. Hybrid work: 3 days in office, 2 remote, with negotiable core hours. Location: Nishi-Shinjuku, Tokyo.