Fostering Practical Skills by Integrating AI Expertise Based on Real-World Engineering Experience

Tokyo, Japan, July 24, 2026—MIXI, Inc. has released its technical training materials for new graduates joining the company in FY2026. This ongoing initiative aims to broadly share the company’s expertise and contribute to the technological advancement of the IT industry. This fiscal year, training programs for 12 subjects will be made publicly available.

In response to the growing use of AI in development work, this year’s technical training expanded the AI-specific program to two days. Furthermore, rather than limiting the focus to AI training, every subject uses real-world AI use cases from current engineers, providing practical learning to help new graduates seamlessly incorporate AI into their everyday development work.

The overviews and materials from the technical training are available below. (Some portions are excluded.)
Training Videos: MIXI TECH&DESIGN (YouTube)
URL: https://www.youtube.com/playlist?list=PLA56pNPDYf8Y
Training Materials:
URL: https://speakerdeck.com/mixi_engineers

About the FY2026 New Graduate Technical Training

Under our purpose of “Enriching communication and inspiring moments of joy,” MIXI operates services across a wide range of business areas. This business areas include sports, lifestyle, digital entertainment, and investment, which means our engineers work with a highly diverse set of technologies. To address this, our technical training for new graduates is conducted in a hands-on, practical format, allowing them to broadly acquire skills in fields such as databases, AI, security, and team development.

In this year’s technical training, we expanded the AI training to two days. In addition to the machine learning (ML) basics covered last year, we introduced more practical topics such as large language models (LLMs) and AI agents. The first day’s theme was “From ML Basics to Operations,” and participants learned the fundamental structure of ML, gained hands-on experience with inference, training, acceleration, and deployment. They also studied common terminology for evaluation and operations. The second day focused on “LLMs, Agents, and Generative AI,” where participants learned the basics of generative AI, practiced working with LLMs, and built their own AI agents. By adding new hands-on sessions on generative AI and AI agents, the program goes beyond theoretical understanding to help new graduates build a foundation for real-world application.

Furthermore, rather than limiting this focus to the AI training itself, every subject integrated real-world AI use cases from current engineers, providing practical learning to help new graduates seamlessly incorporate AI into their everyday development work.

Through this training program designed around the practical application of AI, we aim to nurture next-generation engineers who can drive user value and business results.

The training program for this fiscal year (12 subjects) is as follows.

Comment from Junpei Yoshino, Corporate Officer and CTO of MIXI, Inc.

We will continue to contribute to improving the industry’s overall technical capabilities and developing the next generation of engineers through a range of initiatives, including the external publication of our technical training materials.