Interest
- Post-training Multimodal Foundation Models
- Co-evolution of RL agent and environment for RSI
- Agent Safety and Alignment
Selected Publications
Work Experience
- Applied Scientist @ Woven by Toyota
- Post-training (SFT and RLVR) of VLM for video understanding and visual grounding
- Ultra-long context agentic memory systems for wearable camera
- AI/LLM Engineer @ AnyMind
- Multi-agentic systems for business workflow automation
- Data Scientist @ Travelers Insurance
- Multimodal DL classification with an attention-based interpretability pipeline
- ML Research Scientist @ Hanjuku Kaso
- Offline RL for new actions (TMLR 2026, NeurIPS/ICML/ICLR J2C)
- Offline RL for non-stationarity (SIGKDD 2025)
- ML Research Scientist @ CyberAgent AI Lab
- Offline RL for combinatorial large action space (RecSys 2024)
Education
- Yale University, MS in Statistics and Data Science
- Received NSK Scholarship (USD 70,000 per year, Top 4 in Japan)
- Waseda University, BA in Economics and Certificate in Data Science
- Yale College, Visiting Student in Math and CS
- Top 1 in the grad school courses on causal inference and RL
