Because I Had To

I quit a safe job to build something that didn't exist yet. Not because it was smart. Because I had no other choice.

I've done it three times now.

The Itch (1990s–2012)

I started my career as an entrepreneur — built a web agency in my 20's and sold it in 1999. From 2000–2012 I worked for other people, always with side projects simmering, waiting for the right moment.

In 2012, I read The Lean Startup and Running Lean, discovered companies like Science and BetaWorks. Startup fever hit hard.^[1] Customer-development thinking like The Four Steps to the Epiphany shaped how I saw early-stage building.^[2]

I quit my job, because I had to.

Differential, then the Spinout (2012–2016)

I co-founded Differential, a great growing company with a stellar reputation. As a world-leading Meteor development shop, it was a pretty cool place to work.

But my goal was to go deep on a product, not services. To that end, we built several product MVPs inside Differential. Eventually an opportunity came by that I couldn't resist: take over an analytics product. It was a hard decision to leave the safety of the mothership.

I really, really wanted to build a product company. So I spun out because I had to.

The analytics company was very hard. We battled technical debt and lukewarm investor interest. Onboarding cycles were painful — companies don't like sharing customer data with third-party analytics.

After a gut-wrenching team meeting (pivot vs. persevere is all gut), we pivoted to focus on helping companies "own their data" and share it on their own terms.

That became Astronomer. We built it because we had to.

Astronomer (2016–2022)

The next six years were the hardest and best stretch of my career.

We almost died several times. Each time we found a way through — sometimes by running leaner, sometimes by raising on a thinner deck than I'd have liked, sometimes by sheer luck. Survival was the recurring miracle.

Two decisions mattered more than any others.

The first was bringing on Maxime Beauchemin — the creator of Apache Airflow — as an advisor. That relationship reframed what Astronomer could be. We weren't a generic data-ops company anymore; we were the company best positioned to commercialize Airflow.

The second was hiring Joe Otto as CEO #2. Founder-CEO transitions break a lot of companies. Ours didn't. Joe took the reins and the company kept compounding.

By 2022, Astronomer was growing fast and I had a chance to sell shares in a secondary at a great valuation. I took chips off the table and stepped away.

I left because I had to. After seven years, the conviction that had pulled me into the work had completed its arc. The company was in better hands than mine for what came next, and I needed a reset.

The Gap (2022)

For about six months I did nothing in particular. Caught my breath. Watched the world.

Then Ethan at Venrock made me an offer I couldn't refuse.

I got back in the game because I had to.

Tembo (2022–today)

We started Tembo as a Postgres company.^[3] Managed Postgres, extensions, an opinionated stack — a real bet on the database we all already loved.

Then the agentic age arrived. Not as a trend. As a maelstrom.

Coding agents went from a curiosity to the most consequential shift in software since the cloud, in roughly eighteen months. We watched enterprises struggle to adopt them — too risky to deploy without controls, too valuable to ignore. The orchestration layer didn't exist.

So we pivoted Tembo into AI coding agent orchestration. Same instinct as the Astronomer pivot: a hard team meeting, an honest read on where the customers actually were, and a decision to chase the bigger problem.

The mission is simple: enterprises shouldn't have to choose between agent leverage and agent risk. Tembo is the layer that makes both possible — I unpacked the full thesis in the Tembo manifesto.

We're not giving up. We're persevering through the most chaotic technology transition any of us has lived through, and we're building because we have to.

---

Three companies. Thirty years. Same answer every time.

— Ry