Research Objectives

Building next-generation intelligent NPCs with personalized behaviors and learning capabilities

Research Directions

Application of reinforcement learning in conversational/behavioral intelligent agents, multi-personality modeling and control mechanisms, RLHF/RLAIF, self-reinforcement and generative data-driven multi-stage training flows, behavioral decision optimization under multimodal inputs

Required Skills

Reinforcement learning, language modeling, dialogue system research/development experience, RLHF/RLAIF/PPO/SFT/multi-personality control/reward modeling technologies, proficient use of PyTorch/JAX, independent project management capabilities

Preferred Qualifications

Experience in generative AI and LLM research and development

Benefits

Monthly salary up to 500,000 yen, sufficient GPU resources, collaboration with leading researchers, independent research exploration, top conference publication opportunities, full-time employment opportunities

Position Overview

Seeking excellent interns in speech generation models and large language model reasoning

Research Directions

Generative Speech-to-Speech model research and training, large language model (LLM) reasoning capability research and optimization, speech/text data generation and preprocessing, model fine-tuning through SFT (Supervised Fine-Tuning) and RL (Reinforcement Learning)

Requirements

Master's or PhD student in Computer Science/Artificial Intelligence, proficient in large language model theory, deep learning, and parallel training frameworks (PyTorch, DeepSpeed, etc.)

Preferred Qualifications

Experience publishing papers at top NLP/Speech conferences, practical experience in large-scale model training/tuning

Benefits & Working Conditions

Monthly salary 400,000 yen (varies by experience), online/offline work available, office: Tokyo Shibuya (transportation allowance provided), languages: Chinese/English acceptable, Japanese not required

Key Responsibilities

Drive innovation in LLM, LMM, and next-generation AI through long-term memory, decision models, and autonomous agents; publish papers at top conferences, obtain patents, and contribute to open-source AI communities; explore AI research frontiers and industry trends

Basic Requirements

Solid theoretical foundation in ML, DL, NLP, CV, RL and other fields; Python and C/C++ programming skills in Linux environment; proficiency in technical details of Transformers, multimodal models, generative models, etc.; excellent problem analysis skills; good collaboration attitude and communication skills

Preferred Qualifications

PhD degree in CS, AI-related fields from prestigious universities; experience publishing papers at top conferences/journals; excellent academic competition results (ACM/ICPC, NOI/IOI, etc.); experience participating in well-known large-scale model open-source projects

Key Responsibilities

Obtain large-scale training data for frontier exploration results, build reinforcement learning environments, optimize training efficiency, provide comprehensive automated evaluation systems for next-generation models, apply theoretical breakthroughs to real-world product challenges

Basic Requirements

Python and C/C++ programming in Linux environment, proficiency in PyTorch, large-scale data preprocessing, data generation, and data augmentation capabilities, proficiency in large-scale model training processes, distributed training, and model parallelism, excellent problem analysis skills, good collaboration attitude and communication skills