AI Engineering Summit Tokyo 2026 winter: registration open
TL;DR
- The winter edition of AI Engineering Summit Tokyo 2026 runs Monday, December 14 and Tuesday, December 15 at Hamamatsucho Convention Hall in Tokyo. Findy organizes it, and attendance is free. The registration page opened three ticket types (Day 1, Day 2, both days) at 10:00 on October 22, each capped at 30 people
- Day 1 targets engineers at startups and tech companies, Day 2 engineers at large enterprises. Day 1's English sessions feature Chip Huyen and speakers from Vercel, Duolingo, Uber and Meta. The two GitHub slots have no titles yet
- Only the Japanese sessions are streamed, so the English sessions are in-person only. Interpretation for the English sessions and the schedule for the Japanese sessions have not been announced
- An edition with the same name ran in June and drew 4,708 registrants. June's keynotes came from Databricks, Microsoft and OpenAI. In the winter edition, three of the five English sessions with published titles are case studies from companies using AI in their own development
Contents (6)
Findy, a Japanese engineering platform company, is holding the winter edition of AI Engineering Summit Tokyo 2026 on December 14 and 15 in Hamamatsucho, Tokyo. I’m planning to go. Attendance is free, and registration on the Luma page opened at 10:00 on October 22. Before picking a day, I wanted to know three things: who speaks on which day, what I can watch if I can’t make it to the venue, and how this edition differs from the one in June. I checked the Japanese and English versions of the official site, the registration page and the organizer’s announcements.
Dates and attendance terms
The official site and the registration page list these terms.
| Item | Details |
|---|---|
| Dates | Monday, December 14 and Tuesday, December 15 |
| Hours (registration page) | 9:30 on the 14th to 18:30 on the 15th (Japan time) |
| Venue | Hamamatsucho Convention Hall & Hybrid Studio (5th and 6th floors, Nippon Life Hamamatsucho Crea Tower) |
| Fee | Free |
| Ticket types | Day 1 only, Day 2 only, or both days. All open from 10:00 on October 22, each capped at 30 people |
| Streaming | Japanese sessions only |
| Languages | Japanese and English |
| Lunch | Provided on both days (while supplies last) |
| Organizer | Findy Inc. |
Registration runs through two places. The “Apply” button on the Japanese site leads to Findy’s conference page, while the English site links to Luma. As of October 25 (Japan time), every ticket type on Luma was still open, with 28 spots left for Day 1, 29 for Day 2 and 19 for both days. The page doesn’t say whether the Luma cap matches the venue capacity or is counted separately from Findy’s page.
The sponsor recruitment announcement describes the event as a hybrid of online and in-person. According to the registration page, though, only the Japanese sessions are streamed. Chip Huyen’s talk and the other English sessions are in-person only.
Day 1 and Day 2 target different attendees
This edition splits its audience by day. In the topics section of the official site, Day 1 is “for engineers working on development at startups and tech companies” and Day 2 is “for engineers working on development at large enterprises.”
| Day | Topics listed on the official site |
|---|---|
| Day 1 (14th) | The frontier of AI-driven development, AI product development, agent evaluation, AI-native organizations, token management |
| Day 2 (15th) | Getting started with AI-driven development, building and running AI agents, security and governance, ROI of AI adoption, company-wide rollout and adoption |
The Luma registration page lists the same five topics for Day 2 as for Day 1, which conflicts with the Japanese official site. The Japanese site’s topics are the ones that fit the description of Day 2 as the enterprise day.
If you can only attend one day, the choice probably comes down to whether your team is building something new or rolling AI out across an existing organization. Day 2’s topics, such as security and governance and adoption ROI, are what someone making the internal case for adoption would want to hear.
Announced speakers
The speakers section of the official site lists seven people from overseas companies and five from three Japanese companies. The English site publishes a schedule for the English sessions only.
| Time (Day 1) | Room | Speaker | Title |
|---|---|---|---|
| 9:30 | A | Chip Huyen | From AI Engineering to Agent Engineering |
| 11:30 | A | Lindsey Simon (Vercel) | Building Self-Evolving Apps and Agents |
| 11:30 | B | Kush Agrawal (Duolingo, presenting remotely) | How AI Agents Tripled Our Test Coverage on a 1.8M-Line iOS Codebase |
| 13:30 | A | Mykolas Krupauskas (Uber) | Turning Coding Agents into Production Infrastructure |
| 16:00 | A | Steven Ong (Meta) | Accelerating Agent Observability through Semantic Normalization |
| 17:40 | A | GitHub | Not announced |
For Day 2, the only English session listed is a GitHub slot at 9:30, with no title yet. The speakers section lists two people from GitHub: Idan Gazit, who leads GitHub Next, and Jason Etcovitch, who works on Copilot’s agents. It doesn’t say who takes which slot.
Chip Huyen wrote AI Engineering, and O’Reilly Japan publishes the Japanese edition. Its description says the book covers RAG, agent design, evaluation and guardrails. The official site lists her affiliation as “Stealth” and her title as “Builder.”
The Japanese speakers are two from the Data Strategy Promotion Department at Murata Manufacturing, one from the AI Platform Division at NEC, and two from Tokio Marine & Nichido Systems. The Japanese session schedule isn’t out yet, so it’s unclear which day they speak.
Of the five people speaking in English on Day 1, the talks from Duolingo, Uber and Meta are case studies of building coding agents or agent observability into their own development systems. They look like a good fit for anyone who wants to move from experimenting with agents to making them part of how a team builds software.
How it differs from the June edition
AI Engineering Summit Tokyo also ran in December 2025 and June 2026. The June 8–9 edition also carried the “2026” name and used the same venue. The organizer’s announcements give these registrant counts and keynotes for each edition.
| Edition | Registrants | Keynotes |
|---|---|---|
| AI Engineering Summit (first edition) | 2,742 | Adam Cheyer, Yuki Matsumoto (LayerX) |
| December 2025 edition | 2,828 | Lee Robinson (Cursor), Andrei Eternal (Cline) |
| June 2026 edition | 4,708 | Four slots from Databricks, Microsoft and OpenAI Japan (two from OpenAI Japan) |
| December 2026 edition | Not announced (registration open) | Not announced (Chip Huyen opens Day 1) |
June’s keynotes came from Databricks, Microsoft and OpenAI, companies that provide AI platforms and products. The winter edition’s theme is “creating value with AI,” and three of the five English sessions with published titles are case studies of Uber, Meta and Duolingo using AI in their own development.
After the June edition ended, its official site added a link to watch the archive. The winter site doesn’t yet say whether the same will happen.
Open call for Day 2 speakers closes October 30
The official site is recruiting speakers for Day 2. The themes are “AI agent building and operations at in-house development organizations / AI product development / AI promotion initiatives.” Talks are 10 minutes, and the deadline is 23:59 on Friday, October 30. Submissions go through a Google Form.
The June edition had three open-call slots per day. The winter site doesn’t say how many speakers it will accept. For the December 2025 edition, SoftBank’s event report (Japanese) says three open-call talks were given.
What’s still unknown
I couldn’t find the following anywhere on the Japanese or English official site, the registration page or the organizer’s announcements. I couldn’t fetch the page behind the Japanese site’s “Apply” button, so I haven’t read it.
- The Japanese session schedule, and which day each Japanese speaker presents
- Whether the English sessions will have interpretation or subtitles
- The venue capacity, and whether there will be a lottery if demand is high
- Whether the winter edition will publish an archive afterward
The registration page says the schedule is “coming soon.” Whether interpretation is offered will likely decide whether I go to the venue for the English sessions. I’ll check the official site again once the schedule is out.


