The Hartford · 20262 min read
Building ATLAS, the developer platform for 4,000+ engineers
ATLAS puts onboarding, builds, policy, telemetry, and AI tooling for The Hartford's engineers in one place: a CLI, a desktop agent, and IDE extensions. I build across all three, plus the telemetry behind them.
- Role
- Platform Engineer, Tech & Ops Leadership Development Program
- When
- January 2026 – present
- 4,000+Engineers on the platform (team figure)
- 95,000+Pipelines the platform supports (team figure)
- 144Of my pull requests merged, across 21 repositories
- 80%+Coverage threshold on the tooling I test
The short version
- Problem
- A new engineer's tools were found in a browser, installed from a wiki, and run in a terminal, and nothing connected the three.
- What I did
- Built CLI features, the desktop agent's bridge to the platform site, VS Code extension features, and telemetry, and put cross-platform tests and release checks around them.
- Result
- The team's unified CLI, desktop app, and IDE extensions reached 1,000+ engineers in 90 days; the CLI now runs on 4,000+ machines.
Platform site
Where engineers start
Desktop agent
Runs native commands from the site
HIG CLI
Setup, builds, policy
VS Code extension
The platform in the editor
Telemetry
OpenTelemetry → EventBridge → DynamoDB
Team context
One platform team, one platform
I'm on The Hartford's developer platform team as part of the Technology & Operations Leadership Development Program. The team is led by Anthony Brignano, who sets the technical direction for ATLAS; I'm one of the engineers building it.
As of November 2026, the platform serves 4,000+ engineers, on its way to the full 8,000-engineer organization, supports 95,000+ pipelines, and tracks 13,000+ repositories. Those are the platform's numbers, not mine alone; the sections below are the parts I built.
The problem
Three tools that didn't know about each other
Before ATLAS, an engineer's tooling was discovered in a browser, installed from a wiki, and run in a terminal, and nothing connected the three. Every team's pipelines had their own templates, so one change had to be repeated across thousands of them.
My contribution
The CLI and the desktop agent
The CLI is the core of ATLAS, and the desktop agent is what lets the platform's website reach it.
- Built features in the HIG CLI, now in use on 4,000+ machines.
- Worked on the desktop agent's bridge to the platform site: a button on the site runs a native command on the engineer's own machine, which is what makes the in-browser onboarding possible.
- Fixed bugs and hardened both after rollout.
My contribution
The VS Code extension
The extension brings the platform into the editor engineers already use.
- Built features in the platform's VS Code extension.
- Built the day-one walkthrough that opens VS Code, installs the extension, and tours the parts of the editor that matter on the first day.
My contribution
Telemetry for DevOps Intelligence
The platform has been instrumented since day one, so the team can see how its tools are used. I worked on the telemetry pipeline behind that: events from the CLI, IDE extensions, and desktop agent flow through OpenTelemetry and Amazon EventBridge into DynamoDB, which feeds the usage dashboards.
My contribution
Tests and releases
Across 21 repositories I authored 214 pull requests, 144 of them merged.
- Cross-platform integration tests and an 80%+ code-coverage threshold for the shared developer tooling.
- Release-readiness checks, so the CLI, agent, and extension ship in step.
- Build-farm smoke testing in Harness, so platform changes are checked before engineers depend on them.
Team outcome
How it was adopted
The unified CLI, desktop app, and IDE extensions reached 1,000+ active engineers in their first 90 days. The desktop agent reached the same number in a day, because it shipped to everyone who already had the CLI.
Related
Built with
- HIG CLI
- Desktop agent
- VS Code extension
- OpenTelemetry
- Amazon EventBridge
- DynamoDB
- Harness
- Integration testing