Nuvetralise Book a demo
Book a demo
Studio · est. March 2021

A small studio that takes the pager seriously.

Eleven engineers in a converted brick warehouse on King West. We build one product, work with around forty Canadian fleets, and try not to outgrow our own runbook.

Origin

From a Friday night
page that should have
been a Tuesday email.

Mara and Devon met at Shopify in 2017, working on point-of-sale infrastructure. The company shipped tens of thousands of card readers across Canada, and the on-call rotation was — for a while — a punishment everyone took turns with.

One Friday in 2020, Mara got paged at 23:47 because a metric was 2% above its threshold. The kiosk in question was in a Sobeys in Antigonish that had closed at 21:00. The page was useless. The pattern was not.

Nuvetralise spun out the next March. The first customer was a regional credit union in Saskatchewan with 1,200 ATMs and a long-suffering ops lead named Praveen, who paid an invoice on a handshake and is, as of this month, still on the platform.

Today we are eleven people. We are still small on purpose. The waiting list is real, the deck is short, and most decisions are made over a whiteboard at 460 King West.

2021

Two laptops, one customer

Incorporated in Toronto. First version of the Rust collector shipped six weeks later, on a Sunday afternoon, deployed to 1,200 ATMs in Regina by Monday morning.

2022

YYZ-1 region

Moved off shared hosting onto a dedicated three-node ClickHouse cluster in a Toronto datacentre. Hired our first SRE, Kim, who still owns the rotation.

2024

Anomaly engine, Mark IV

Threw out the third rewrite. The fourth one, written in Python in eight weeks, is the one we still ship. It does less. That's the point.

2026

Forty fleets, eleven people

Approaching 65,000 devices on stream. Still no salespeople. Still answering the support inbox in rotation. We'll see how long that lasts.

Five things we believe
about quiet software.

Strong opinions, occasionally revised. Last edited March 2026.

Principle 01

A page is a contract.

If we wake an engineer up, we are saying: there is something only you can do, right now. Anything less than that is a dashboard, not an alert.

Principle 02

Defaults are a moral choice.

The defaults you ship are the system 90% of customers will run forever. If they're wrong, the spreadsheet of regret is long. We argue about defaults more than features.

Principle 03

Boring stack, weird culture.

Postgres, Go, ClickHouse, Rust. The novelty budget goes into product, not infrastructure. We are, however, the only team in Toronto with a Friday "delete a feature" ritual.

Principle 04

Customers, not users.

The word "user" makes us write worse software. We know our customers' first names, their fleets, what's on their wall. The product reflects that — for better or worse.

Principle 05

Stay hireable.

We pick technology that, if Nuvetralise disappeared on a Tuesday, you could keep running until you found a replacement. No bus-factor of one. No magic.

Principle 06

Eat your own pager.

Every founder is in the on-call rotation. Every quarter. No exceptions. It is the only way we know to stay honest about what shipping a feature actually costs the people who maintain it.

The studio

King West, fourth floor.

We share a brick-and-bowtruss floor with a small architecture practice and a coffee roaster who supplies the office. There is a printer that hasn't worked since November. Two whiteboards. A view of the rail corridor.

If you're in town and curious — coffee is real, the door buzzer is sometimes broken, and we'll show you the half-finished v2 of the console that nobody outside the building has seen.

Drop a note to hello@nuvetralise.com first. We don't do walk-ins, mostly because the elevator is moody.

Quiet workspace with a notebook, mechanical keyboard and brick wall in soft daylight