​Description

​What does it take to make enterprise AI useful, secure, and deployable in real organizations? Join Tokyo AI and Cohere for an evening with the teams building and deploying Cohere's North platform and AI models.

​Aidan Gomez opens the programme in conversation with Ilya Kulyatin. Cohere's Japan and technical teams then share product demonstrations and practical lessons on agent workflows, security, forward-deployed engineering, customer infrastructure, and model customization. The evening is for software and AI engineers, architects, infrastructure teams, and enterprise technology leaders who want to understand how these systems work in practice.

​When: Thursday, November 12, 2026, 18:00-21:00 JST
Where: Tokyo, venue to be announced. The address will be shared with approved guests.

​Agenda

​The talks are in English. The programme begins at 18:00.

• ​18:00-18:30 - Talk 1: Aidan Gomez in conversation with Ilya Kulyatin

• ​18:30-18:40 - Talk 2: Own Your AI: A Brief Introduction to Cohere, Its Products and Customers (Takeo Tokunari)

• ​18:40-18:55 - Talk 3: Making Sovereign AI a Tangible Product (Yuta Kihara)

• ​18:55-19:05 - Talk 4: Vulnerability Management in the AI Era: Lessons from Practice (Jack Zhang)

• ​19:05-19:15 - Audience Q&A

• ​19:15-19:30 - Break

• ​19:30-19:40 - Talk 5: Technical Deep Dive: How Forward-Deployed Engineers Build Custom Agents (Taiga Kubota)

• ​19:40-19:50 - Talk 6: Technical Deep Dive: Running Cohere in Your Environment (Connor McBrine-Ellis)

• ​19:50-20:05 - Talk 7: Technical Deep Dive: Cohere’s Generative, Retrieval and Custom Models (Steven Connors)

• ​20:05-20:20 - Audience Q&A, facilitated by Takeo Tokunari

• ​20:20-21:00 - Networking with food and drinks

​Talks

​Talk 1: Aidan Gomez in conversation with Ilya Kulyatin

​Cohere’s co-founder and CEO, Aidan Gomez, joins Tokyo AI founder Ilya Kulyatin for a fireside chat. Aidan co-authored “Attention Is All You Need,” the paper that introduced the Transformer.

​Talk 2: Own Your AI: A Brief Introduction to Cohere, Its Products and Customers (Takeo Tokunari)

​Would you send a sensitive prompt to a cloud model? What would it take to maintain your own retrieval pipeline and agents over time? Takeo Tokunari, Customer Success Manager at Cohere, introduces the choices facing enterprise teams in Japan. He draws on Cohere’s work with Japanese customers, including Fujitsu and Mitsubishi Heavy Industries.

​Talk 3: Making Sovereign AI a Tangible Product (Yuta Kihara)

​Yuta Kihara, Solutions Architect at Cohere, shows what sovereign AI looks like on screen. He will walk through North 2.0, which organizations can run in their own environment, and demonstrate a custom agent for a government use case and an automated manufacturing workflow. Along the way, he will show how builders evaluate agents and how teams govern their use.

​Talk 4: Vulnerability Management in the AI Era: Lessons from Practice (Jack Zhang)

​A new software vulnerability is disclosed. How do you find out which deployments are affected and what to fix first? Jack Zhang, Engineering Program Manager at Cohere, explains how his team uses AI to investigate reported vulnerabilities, assess their impact with customers and partners, and coordinate remediation.

​Talk 5: Technical Deep Dive: How Forward-Deployed Engineers Build Custom Agents (Taiga Kubota)

​If North lets customers build agents without writing code, when do they need an engineer? Taiga Kubota, a forward-deployed engineer on Cohere’s Agents team, walks through an anonymized customer engagement. He explains how the team decides which capabilities belong in North for everyone and which need to be built for one customer.

​Talk 6: Technical Deep Dive: Running Cohere in Your Environment (Connor McBrine-Ellis)

​For a government or regulated enterprise, even sending a prompt to a shared API can be off the table. Connor McBrine-Ellis, from Cohere’s forward-deployed Infrastructure team, explains how GPU choice, model performance, and permissions shape a deployment. He will also cover the engineering challenges of running the system locally on a network isolated from the internet.

​Talk 7: Technical Deep Dive: Cohere’s Generative, Retrieval and Custom Models (Steven Connors)

​Steven Connors, from Cohere’s FDE Agents team, explains how Cohere adapts generative and retrieval models for individual customers and languages. He will discuss North-Small-Translate, Cohere’s publicly released translation model, and draw on the team’s experience training custom Japanese-language models.

​Organizers

​Tokyo AI (TAI) and Cohere organize this event together.

​Ilya Kulyatin is an entrepreneur with work and academic experience in the US, Netherlands, Singapore, UK, and Japan. He holds a BA in Economics, an MA in Finance, and an MSc in Machine Learning. He is a three-time founder, building the AI community in Japan through Tokyo AI while leading Foundry Labs株式会社.

​Supporters

​Cohere is an enterprise AI company building foundation models and end-to-end products for business use. Its North platform brings together AI agents, enterprise knowledge, automation, and governance to help organizations put AI to work. With deployment options across cloud, private cloud, and on-premises environments, Cohere gives enterprises control over their data, security, and AI systems.

​Foundry Labs K.K. is a Tokyo-based AI systems integrator and solutions provider, delivering end-to-end support for enterprises from strategy through implementation, deployment, and operations. Its work is tailored to clients' operational, regulatory, and security needs.

​About TAI

​Tokyo AI (TAI) connects the people building AI in Japan with the global ecosystem through technical events and community collaborations.

​Privacy Policy

​We will process your email address for event-related communications and ongoing newsletter communications. You may unsubscribe from the newsletter at any time. Further details are available in our Privacy Policy.