Research Engineer, LLM modeling

Role & Expectations

1. 1Conceive, develop, and compare different agent harnesses (memory, context compression, inter-agent communication architectures, etc.)
2. 2Design and implement rigorous quantitative benchmarks for large-scale agentic tasks
3. 3Support automated evaluation of models and prompts, ensuring quality across the entire lifecycle — from training to production
4. 4Collaborate with the product organization to solve the hardest challenges in applying agents to products
5. 5Create and optimize training data mixes to improve agent task performance and usability
6. 6Transfer research outcomes to the Agentic Product Engineer team, raising quality across all products

Job Description

1. 1Agent Research & Development: conceive, develop, and compare agent harnesses; research reasoning, planning, retrieval; develop multimodal and long-context technologies; survey and improve research papers
2. 2Evaluation & Benchmarking: design quantitative benchmarks for agentic tasks; design synthetic data generation and evaluation benchmarks; support automated evaluation of models and prompts
3. 3Production Problem-Solving: optimize inference latency and cost (quantization, distillation, caching); create and optimize training data mixes; advance agent evaluation frameworks; improve quality in production
4. 4Knowledge Transfer & Outreach: transfer technology and mentor Agentic Product Engineer team; collaborate with academic institutions and OSS communities

Why We're Hiring

1. 1JAPAN AI STUDIO aims to be 'the brain of the enterprise' integrating SaaS tools for autonomous AI agents
2. 2Frontier challenges include reasoning quality limits, long-term memory design, multimodal handling, and low-latency inference
3. 3Strengthening team to build more effective agents for long-horizon tasks, design coordination mechanisms, and solve harness/infrastructure/fine-tuning challenges

Mission

1. 1Solve problems that make today's agents give up
2. 2Break through quality limits of reasoning, retrieval/planning, long-term memory, and tool use
3. 3Pave the way for hundreds of workflows on JAPAN AI STUDIO to operate smarter, faster, and safer

Why You'll Love This Role

1. 1Research powers 'the brain of the enterprise' — not just chatbots
2. 2Research directly connects to production used by ~200 companies
3. 3Work on frontier challenges: reasoning quality, long-term memory, multi-agent coordination
4. 4Encourages paper publication and tech blog writing, supports academic collaboration
5. 5Transfer methods to Agentic Product Engineer team and exercise leadership
6. 6Rapid-growth startup with 200+ people and 9 products in 3 years, significant autonomy

Key Results (KR/Metrics)

1. 1Benchmark score improvement rate (internal and public benchmarks)
2. 2Number of novel methods shipped to production (per quarter)
3. 3Inference latency and cost reduction rate
4. 4Number of papers and technical blog posts published
5. 5Number of internal knowledge transfers completed

Team Structure

1. 1Approximately 120 members in development organization
2. 2Research Engineers work across JAI Lab, AI & Model, Voice & Tel groups
3. 3Closely collaborating roles: Agentic Product Engineer, Agent Harness Engineer, AI Quality Scientist, Product Manager

Learning & Development Support

1. 1Company covers AI tool usage costs (JAPAN AI SaaS, Cursor, ChatGPT, Claude)
2. 2Development tool support up to ¥30,000 per year
3. 3Book purchase assistance up to ¥30,000 per half-year
4. 4Language learning and qualification support
5. 5Refresh allowance up to ¥5,000 per month
6. 6Housing allowance up to ¥30,000 per month for designated areas

応募資格（必須）

• Master's or Ph.D. in Computer Science, Software Engineering, Artificial Intelligence, Machine Learning, Mathematics, Physics, or related fields
• Experience developing complex agentic systems using LLMs
• Significant hands-on experience in software engineering and ML
• Experience with LLM prompt engineering and/or building products with language models
• Experience with large-scale model training and inference using PyTorch or JAX
• Deep understanding of LLM and Transformer architectures
• Ability to read, reproduce, and improve upon research papers
• Strong implementation skills in Python (production-quality code)
• Language requirement (at least one of the following): Japanese: Fluent — able to discuss product development without friction; English: Business level

歓迎する経験

• Experience with large-scale reinforcement learning on language models
• Experience designing and implementing multi-agent systems
• Publications at top-tier conferences (NeurIPS, ICML, ACL, EMNLP, or equivalent)
• Hands-on experience implementing alignment techniques such as RLHF or DPO
• Experience with multimodal models (e.g., Vision-Language models)
• Background in agent evaluation or AI safety research
• Ph.D. in CS, ML, NLP, or a related field
• Ability to communicate research findings in English

待遇・福利厚生

• AIツール利用費補助
• Development Tool Support
• Language Learning / Qualification Support
• Refresh Allowance
• シャッフルランチ/ディナー
• 夏季休暇
• 年末年始休暇
• 年次有給休暇
• 従業員持株会
• 慶弔休暇
• 社会保険完備
• 資格取得・英語学習支援
• 部活動手当

求めるスキル・経験