Highreso and APTO Co-Host Webinar on "The Formula for Successful LLM Data Development" Supporting Physical AI
Highreso Inc., which operates the GPU cloud service "GPUSOROBAN," will co-host a webinar with APTO Inc. on the theme of "The Formula for Successful LLM Data Development." This webinar will cover LLM safety design, high-quality data development, and its potential for physical AI, as well as the requirements for computing infrastructure and GPU environment design for optimizing development speed and cost.
📋 Article Processing Timeline
- 📰 Published: April 1, 2026 at 19:01
- 🔍 Collected: April 1, 2026 at 10:15
- 🤖 AI Analyzed: April 22, 2026 at 03:29 (497h 13m after Collected)
Highreso Inc. (Headquarters: Shinjuku-ku, Tokyo; Representative Director: Yoshiyuki Shikura; hereinafter "the Company"), which operates the GPU cloud service "GPUSOROBAN," will co-host a webinar with APTO Inc. on the theme of "The Formula for Successful LLM Data Development."
### Webinar Overview
This seminar will organize topics ranging from the premise of social implementation, which is LLM safety design, to the concept of high-quality data that determines performance, and further to the potential for deployment to physical AI.
Furthermore, it will explain the indispensable requirements for "computing infrastructure" to put these design philosophies into practice, and share perspectives on GPU environment design to achieve both development speed and cost optimization.
・Date: Thursday, April 23, 2026, 14:00-15:00
・Viewing method: Online (Zoom)
・Capacity: 1000 people
・Participation fee: Free
・Application here: https://soroban.highreso.jp/423webinar
### Target Audience
・Those who want to systematically understand LLM safety design and the concept of high-quality data
・Those who are working on improving LLM accuracy in fields such as medical, instruction following, and mathematics
・Those who want to grasp the latest trends in dataset design required for Physical AI, both domestically and internationally
・Those interested in designing and optimizing GPU infrastructure to accelerate AI development
### Speakers
・Yoichi Kano, APTO Inc.
After working at a general publishing company and a cybersecurity company, joined APTO in 2022. Assumed current position in 2024. Continues to support companies in solving AI challenges starting from data. Currently, in addition to overseeing business and corporate sides, actively develops and publishes research data with domestic and international research institutes, universities, and operating companies.
・Takuya Omi, APTO Inc.
While enrolled in graduate school, engaged in planning work for news media, involved in business efficiency improvement using natural language processing libraries, AI system development, and direction of product improvement based on user behavior data. Subsequently, as a UX/UI designer for a video streaming service, was responsible for user experience design and data-driven problem analysis.
Currently, as a project manager overseeing service planning and development utilizing AI, is engaged in designing and promoting new products applying natural language processing (NLP) and large language models (LLM). Based on LLM learning experience, also responsible for creating synthetic data and designing learning methods that lead to accuracy improvement.
・Shunsaku Endo, APTO Inc.
Began seriously learning programming while enrolled in a game programming vocational school, and received numerous awards at in-school judging committees during his enrollment. Developed a strong interest in the AI field while honing his technical skills, and co-founded APTO in 2020. Focused on the challenge of inefficiency in data collection, preparation, and utilization in AI development, and is working to solve it. Currently at APTO
### Webinar Overview
This seminar will organize topics ranging from the premise of social implementation, which is LLM safety design, to the concept of high-quality data that determines performance, and further to the potential for deployment to physical AI.
Furthermore, it will explain the indispensable requirements for "computing infrastructure" to put these design philosophies into practice, and share perspectives on GPU environment design to achieve both development speed and cost optimization.
・Date: Thursday, April 23, 2026, 14:00-15:00
・Viewing method: Online (Zoom)
・Capacity: 1000 people
・Participation fee: Free
・Application here: https://soroban.highreso.jp/423webinar
### Target Audience
・Those who want to systematically understand LLM safety design and the concept of high-quality data
・Those who are working on improving LLM accuracy in fields such as medical, instruction following, and mathematics
・Those who want to grasp the latest trends in dataset design required for Physical AI, both domestically and internationally
・Those interested in designing and optimizing GPU infrastructure to accelerate AI development
### Speakers
・Yoichi Kano, APTO Inc.
After working at a general publishing company and a cybersecurity company, joined APTO in 2022. Assumed current position in 2024. Continues to support companies in solving AI challenges starting from data. Currently, in addition to overseeing business and corporate sides, actively develops and publishes research data with domestic and international research institutes, universities, and operating companies.
・Takuya Omi, APTO Inc.
While enrolled in graduate school, engaged in planning work for news media, involved in business efficiency improvement using natural language processing libraries, AI system development, and direction of product improvement based on user behavior data. Subsequently, as a UX/UI designer for a video streaming service, was responsible for user experience design and data-driven problem analysis.
Currently, as a project manager overseeing service planning and development utilizing AI, is engaged in designing and promoting new products applying natural language processing (NLP) and large language models (LLM). Based on LLM learning experience, also responsible for creating synthetic data and designing learning methods that lead to accuracy improvement.
・Shunsaku Endo, APTO Inc.
Began seriously learning programming while enrolled in a game programming vocational school, and received numerous awards at in-school judging committees during his enrollment. Developed a strong interest in the AI field while honing his technical skills, and co-founded APTO in 2020. Focused on the challenge of inefficiency in data collection, preparation, and utilization in AI development, and is working to solve it. Currently at APTO