Sydney
Runtime and platform
The agent harness, the execution environment, the enterprise layer, applied science.
Record 08 · the house
Four years, three cities, one bet.
That the useful unit of AI at work is not a faster answer but a run that finishes. Everything the company has built since 2022 follows from that one position, including the awkward parts of it.
T+00:00The record
| Field | Value |
|---|---|
| Founded | 2022 |
| Head office | Palo Alto, California |
| Engineering | Sydney · Vancouver |
| People | Around thirty-five |
| Consumer users | 500,000+ |
| Backers | Amazon Alexa Fund, SRI Ventures, DCVC, Candou Ventures, Infosys Innovation Fund |
Small, and
that is the
whole point.
Thirty-five people cannot win by having more of everything. They can win by picking one hard thing, which here is making a run survive a week without a person, and by letting a services firm with a hundred thousand consultants do the deployments.
T+00:32Who builds it
A Google
product half
and an AWS
systems half.
The founders had already built and sold companies before this one. The engineering bench came, largely, out of one building in Sydney.
Co-founder
Chief executive
Berkeley and Stanford, founded a recommendation company sold to Google in 2011, then eleven years inside Google in senior product leadership on analytics and measurement.
Co-founder
Product and business
Illinois, five years leading AI work at Meta, previously founded and sold a video-intelligence company.
Co-founder
Chief science officer
Six years as a senior applied scientist at AWS on reinforcement learning for core infrastructure, multiple patents, published research before that.
Engineering
The Sydney bench
Long-tenure AWS engineers in networking, deployment safety and automation. The people who have run systems that are not allowed to fall over.
Roles are described rather than named, because this is a design study and not the company's own page.
T+01:10What is open
Hiring in
Sydney and
the Bay.
The platform work sits in Sydney: agent runtime, the harness the runs execute in, the enterprise layer. The Bay Area seats are the ones that go and stand it up in front of a customer, which at this size is the same job as finding out what to build next.
The live list, with what is actually open today, is on the company's own job board.
Bay Area
Deployment and field
Forward deployed engineering, solutions architecture, applied AI, sales.
T+01:44About this site
This is an unofficial redesign of the NinjaTech AI marketing site, built as a design study by Michel Grolet. It reuses none of the original's structure, sections, wording, imagery or marks. What it keeps is the product: the same platform, the same public facts, and links that go to the real console and the real account flow.
Product facts, client names, plan prices and the partnership are all taken from NinjaTech's own public pages. Anything drawn as a chart is labelled illustrative. No performance benchmark, customer outcome or private commercial term appears anywhere on this site.
| Colophon | |
|---|---|
| Type | Archivo, Martian Mono |
| Ink | #14130F on #F1EDE4 |
| Signal | #D8341F |
| Built | September 2026 |