Tokyo AI Talks (TAIT) #5

​Objective

​​TAIT is a community composed of people based in Tokyo and working with, studying, or investing in AI. We are engineers, product managers, entrepreneurs, academics, and investors intending to build a strong “AI coreˮ in Tokyo. This core is composed of a set of nodes (all of us), and we want to increase the valency of each node through knowledge sharing and mutual connections.

​​Topics

​​In this session we will explore the 3 major pillars of engineering AI models; training, inference, and serving. As the field keeps advancing, we think it is important to help the community keep up to date with the latest methodologies and tools to optimize these key steps in AI engineering. Our focus will encompass a range of topics, including:

• ​Training Large and Small Models: Delve into the complexities of training large language models (LLMs) utilizing high-performance computing (HPC), and examine the techniques for training small models, such as Small Language Models (SLMs) and Mixture of Experts (MoE). Our speakers will share insights and strategies for managing computational demands, optimizing performance, and leveraging the benefits of efficiency and adaptability in various applications.

• ​Inference Speed and Latency: Discuss approaches to enhancing inference speed and reducing latency, particularly through the use of ML compilers. Our speakers will share techniques and tips to successfully compile ML models for speed and accuracy.

• ​Model Serving: Compare and contrast the serving of large models versus edge models, addressing the requirements and challenges of each.

​Spkears selection is ongoing, let us know if you have some candidates we could reach out to!

​​Participant profiles

​​Last time we had 65 participants: 62% engineers, 18% product, 15% investor, and 5% students.

​​Organizers - alphabetic order

​​Ilya Kulyatin: Fintech and AI entrepreneur with work and academic experience in the US, Netherlands, Singapore, UK, and Japan, with an MSc in Machine Learning from UCL.

​​Hiroki Nakayama: AI engineer at Softbank, MSc in Applied Physics from Stanford and Keio. Hiroki is currently working in MLOps and AI Infrastructure, with a keen interest in stimulating the open-source community in Japan.

​​Xiaoli Shen: 10+ years of software engineering and AI solutions architecture in companies like AWS and Fast Retailing, between Germany and Japan, with an academic background in Digital Media (novel interfaces).