The challenge
Garner Health crunches billions of medical-claims records to rank physicians on real patient outcomes, to get more care to the best-performing doctors in the nation, reversing the trend of surging healthcare costs and lower quality care.
That mission drives a heavy Site Reliability Engineering (SRE) workload:
- Complex ETL pipelines that must run 24 × 7.
- Rapid head-count growth – started the year with 40 → ending with over 100.
- Ambitious infrastructure ideas (dynamic resource provisioning, state drift detection) that were too big to prototype in spare cycles.
“As humans, we can only keep so much in working memory,” notes Forrest Thomas, Site Reliability Engineer. “Big systems stall when you can’t hold every moving part at once.”
Why Augment Code?
Garner Health evaluated several AI coding tools, including GitHub Copilot, Cursor, and Windsurf. Augment Code stood out for several technical reasons:
| Requirement | Augment advantage | Forrest's take |
|---|---|---|
| Speed & quality | Sub-second completions and higher-fidelity code | “Cursor felt sluggish; Augment replies land while I’m still in flow.” |
| Security & compliance | No-training option on customer code, Customer-Managed Keys, Okta SSO | “HIPAA compliance was a hard gate—Augment cleared it.” |
| Code-base awareness | Full-project indexing and MCP integration for architectural context | “It actually understood our Go & Python repos.” |
| IDE flexibility | Lightweight extension for JetBrains, VS Code & Vim | “We aren’t forcing devs to abandon their tools of choice.” |
The tipping point? A weekend POC: Augment generated ≈ 10 K lines of production-ready Go for Forrest’s dynamic-provisioning project—something he’d postponed for years.
The results
"Augment fundamentally changed how we approach software development," Forrest concluded. "It has become essential—not just for increasing productivity but for improving the overall quality and design of our systems."
Garner Health’s adoption of Augment produced clear technical advantages:
- Accelerated Development: Developers quickly moved from idea to functional prototypes, significantly enhancing the team's capacity for innovation. "Augment quickly moves you from ideation into a coding flow state," Forrest explained. "I don't have to deliberate on language semantics like 'should I use a for loop or a map?'—it keeps my attention on the bigger problem."
- Higher Code Standards: Augment's precise coding assistance reduced the number of manual interventions, maintaining quality even during rapid team growth.
- Scalable Practices: Augment facilitated consistent coding standards across the expanding team, streamlining onboarding and collaboration. Forrest emphasized, "With Augment guidelines and MCP integration, we can coordinate coding standards at an organizational level."
- Seamless Rollout: 60 developers onboarded in two days; positive feedback and zero IDE switches.
“Augment finally lets me focus on designing elegant systems instead of juggling imports and boilerplate. I can’t imagine coding without it.” — Forrest Thomas, Site Reliability Engineer, Garner Health
A tool for professional software engineering, not vibe coding
“If you only want to hit Tab, you’ll ship a lot of code you don’t understand,” Forrest warns. That “vibe coding” style churns out lines quickly but leaves teams debugging brittle, opaque systems later.
With Augment Code, he says, “I jump from idea straight into flow. Augment holds the whole system in context, so I can focus on architecture and outcomes.” Instead of wrestling with imports or syntax, Forrest applies his expertise to designing extensible, maintainable services.
In short: vibe coding automates keystrokes; Augment automates the grunt work, letting engineers practice real software engineering—thinking deeply about why and how a system should be built.