Software Architect2012-2014
Pearson logo

Pearson

Nibiru: Where My Cloud Foundation Began

Building the automation platform that taught me how to wield AWS—and started a decade-long journey

Minutes
vs Months
Infrastructure Provisioning
Flask UI
Self-Service Portal
Gene Kim
Consulting Engagement
Nibiru
Platform Born

The Directive

In 2012, Pearson's CIO Aref gave a simple directive: "Take us to the cloud." I was a software architect tasked with figuring out how to deploy MongoDB in AWS. What started as solving a single deployment problem became something much bigger.

The existing process was manual and slow—infrastructure provisioning took months. Every deployment was a custom snowflake. I started writing a Python utility called "awsdeploy" to automate the work. The first commit was August 18, 2012.

By December, what had been a simple deployment script had evolved into a full platform. I named it Nibiru—after the mythological planet. This was where I first learned how to wield AWS, and it shaped everything that came after.

Part 1: Building Nibiru

Nibiru wasn't just deployment automation—it was a complete platform for managing cloud infrastructure. The philosophy was radical for the time: give developers what they need, enforce operational best practices through automation, and make the platform do the heavy lifting.

Nibiru Platform Architecture

Python + Flask UI

Self-service deployment portal with REST APIs

Puppet

Centralized configuration management

Zabbix

Centralized monitoring and alerting

Multi-Service Support

MongoDB, Redis, Elasticsearch, and more

One of the key innovations was the naming convention. Every piece of infrastructure followed a predictable pattern like use1a-pri-mongodb-s1-01—region, environment, service, shard, instance. This made the infrastructure self-documenting and enabled automation at every level.

Part 2: Gene Kim and The Phoenix Project

In January 2013, Gene Kim's "The Phoenix Project" launched. Shortly after, Pearson's senior leadership brought him in for a three-day consulting engagement. He walked through the organization's pain points—technical debt, dev and ops collaboration, the importance of version controlling production artifacts.

Gene Kim's Three Ways

1

Flow

Optimize the flow of work from dev to ops to customer

2

Feedback

Amplify feedback loops to fix problems at the source

3

Continual Learning

Create a culture of experimentation and learning from failure

At the time, I didn't realize that what we were building with Nibiru was DevOps. I just thought the cloud itself was enabling us to do things faster. Gene's engagement gave me the vocabulary to understand what we were actually doing—and why it mattered.

Part 3: The Philosophy That Lasted

Nibiru embodied a philosophy that I've carried through every engagement since: give developers what they need, but enforce operational best practices through the platform itself. Developers got root access—but Puppet enforced state. They could deploy anything—but the platform handled the how.

The Pattern Emerges

  • Self-service over tickets
  • Automation over documentation
  • Guardrails over gatekeepers
  • Minutes over months

This wasn't just about technology. I introduced colleagues to the concepts of DevOps and cloud architecture—people who would go on to lead their own transformations. A decade later, one of those connections led me to the Gilead engagement, where I'd apply everything I'd learned at scale.

Key Lessons

Start With a Problem

Nibiru started as a script to deploy MongoDB. Solve real problems and platforms emerge organically.

Relationships Compound

The people you work with become your network. A connection from 2013 led to Gilead in 2022.

Learn the Vocabulary

I was doing DevOps before I knew the word. Gene Kim gave me the language to explain what I'd built.

Further Reading

Starting Your Cloud Journey?

Whether you're building your first platform or scaling an existing one, let's talk about getting it right from the start.

Schedule a Conversation