Applied LLM Researcher

**About the company:** AssetHub Tokyo AssetHub is a next-generation 3D creation tool that uses AI to help creators build and edit 3D assets up to 100× faster. Built for game developers, it changes how games and 3D content are made by bringing together Games, IP, and AI to deliver speed, scale, and globally competitive quality. Read more **Responsibilities:** Conduct research and development on Large Language Models (LLMs) and multimodal models (text, image, 3D) Research and integrate image generation models (e.g., Diffusion, Transformers, NeRF / 3DGS) with LLMs Design AI systems for game development and creative production pipelines Design, implement, and operate LLM-based agent systems, including: Tool use, planning, memory, and multi-agent architectures Stabilization and optimization for real-world production environments Build evaluation frameworks and metrics for LLMs and agents Offline evaluation, online evaluation, and human-in-the-loop assessment Reproduce, extend, and apply ideas from recent research papers Collaborate closely with engineering and product teams to bring research into production Requirements PhD (completed or in progress) in Computer Science, Machine Learning, or a related field Strong background in LLMs or generative models Solid implementation skills in Python and PyTorch (or equivalent frameworks) Experience or strong understanding of bringing research code to production-quality systems Ability to read papers, implement them, and translate research into real products Nice to haves While not specifically required, tell us if you have any of the following. Deep expertise in image generation and multimodal generative models Diffusion models, VAEs, Transformer-based architectures Strong interest in creative domains and games Game development, 3D content, IP creation, UGC platforms, creative tools Hands-on experience with LLM agents in production Designing, deploying, and iterating on agent systems in real environments Practical experience in evaluation and measurement Designing quantitative metrics for LLMs and agents A/B testing, logging, quality monitoring Experience with large-scale distributed training or inference optimization Open-source contributions or research publications (top-tier venues not required) Experience bridging the gap between research and production Ability to define ill-posed problems and drive them to solutions Strong product sense and user-centric thinking, especially for creators Familiarity with game engines (Unity, Unreal) or DCC tools (e.g., Blender) Experience working in startups or small, fast-moving teams APPLY FOR THIS POSITION DO YOU NEED MORE INFO? ASK A QUESTION **Requirements:** PhD (completed or in progress) in Computer Science, Machine Learning, or a related field Strong background in LLMs or generative models Solid implementation skills in Python and PyTorch (or equivalent frameworks) Experience or strong understanding of bringing research code to production-quality systems Ability to read papers, implement them, and translate research into real products **Nice to have:** While not specifically required, tell us if you have any of the following. Deep expertise in image generation and multimodal generative models Diffusion models, VAEs, Transformer-based architectures Strong interest in creative domains and games Game development, 3D content, IP creation, UGC platforms, creative tools Hands-on experience with LLM agents in production Designing, deploying, and iterating on agent systems in real environments Practical experience in evaluation and measurement Designing quantitative metrics for LLMs and agents A/B testing, logging, quality monitoring Experience with large-scale distributed training or inference optimization Open-source contributions or research publications (top-tier venues not required) Experience bridging the gap between research and production Ability to define ill-posed problems and drive them to solutions Strong product sense and user-centric thinking, especially for creators Familiarity with game engines (Unity, Unreal) or DCC tools (e.g., Blender) Experience working in startups or small, fast-moving teams APPLY FOR THIS POSITION DO YOU NEED MORE INFO? ASK A QUESTION