TLDRocket
Sign in

[Event Report] Sakana AI Engineer Open House 2026

Sakana AI

Sakana AI held a second engineer open house, now across five roles instead of one. The pitch is bigger than hiring: it’s building “AI sovereignty” for Japan.

Based on reporting by Sakana AI — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

Sakana AI held its second engineer open house on October 5, 2026, and this year the company widened the tent. Last year’s event focused on Applied Research Engineer roles. This time it covered five: Applied Research Engineer, Forward Deployed Engineer, Software Engineer, Solution Engineer, and Cybersecurity Engineer. The format was simple enough: people from each role explained what they do, what makes the work interesting, and then fielded a lot of questions.

The event opened with cofounder and CEO David Ha, whose own career is part of the company’s pitch. He moved from Goldman Sachs derivatives trading into research, then worked at Google on projects such as World Models and Sketch-RNN. He told attendees that he, too, had changed careers. The company name comes from the picture-book Swimmer, and the underlying idea is collective intelligence: not one fish, but the group. Sakana AI is trying to move in its own direction rather than follow everyone else.

That philosophy shows up in the company’s big theme: AI sovereignty. David argued that overseas open-source models can carry the political and cultural bias of the countries they were built in, which can lead to wrong answers or refusals when asked about Japanese politics or culture. Sakana AI is tackling that across three layers. It is building a native Japanese agentic LLM from scratch in pre-training, adapting foreign models to Japanese customs with Namazu in post-training, and preparing Fugu for orchestration so multiple models can be switched in and out if geopolitical risk makes one unavailable.

The company also said its product lineup has expanded quickly over the past half-year with Sakana Chat, Sakana Marlin, Sakana Fugu, and Sakana Translate. David framed the next phase around enterprise business, government and defense work, and product business, with the goal of becoming a global AI company from Japan. Later, the team leads drilled into the details. The finance-focused Applied team works with major institutions such as MUFG, Daiwa Securities Group, and SMBC Group on decision-support systems for planning, macro analysis, M&A proposals, and advanced credit decisions — problems where many data sources and serious financial consequences are on the line.

Product and GTM got their own time too. Product head Ohmura said the team has released Sakana Chat, Sakana Marlin, Sakana Fugu, Sakana Translate, and Sakana Namazu in about six months, often pulling use cases from Applied projects and ideas from Research. The GTM team, launched only in September 2026, already has three account executives and several forward deployed engineers. Across the afternoon, the message was consistent: Sakana AI is not trying to keep research, product, and delivery in separate boxes. It wants them feeding each other like the company mascot story suggests.

My take — AI-written commentary, not fact-checked reporting

Sakana AI’s real bet is not that Japanese firms need one more chatbot. It’s that the boring, hard plumbing around model routing, adaptation, and delivery matters more than the demo. That is a much healthier instinct than the usual AI theatre, and frankly a relief.

Read more about this at: Sakana AI

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.