🔥 The Plumbers, the Electricians, and the Guardians of the Galaxy

How Investing in 'Invisible' Engineering Disciplines Built a $165M Innovation Engine

In 2015, HealthFusion was acquired by NextGen Healthcare for $165 million. When people asked what made us attractive, the answers were usually about our cloud-based EHR, our elegant user experience, our growing customer base.

But here's what I know: one of our quiet competitive advantages was that we could ship weekly releases with fast recovery. While competitors were stuck in quarterly release cycles—often buggy, often delayed—we were iterating, learning, and improving every single week.

That didn't happen because we had brilliant engineers (though we did). It happened because we invested in the people nobody sees: the DevOps engineer, the DBA, the QA champions, the infrastructure specialists.

We didn't just build a product. We built an innovation engine. And that engine was valuable.

This blog is an ode to those unsung heroes—the DevOps engineers, DBAs, QA champions, and infrastructure specialists. The people who wire automation, watch logs, read the fine print on compliance, and carry risk on their shoulders so others can ship, pivot, demo, and dream.

But it's also a business case: investing in these 'invisible' disciplines isn't overhead. It's how you build an innovation engine that investors will pay a premium for.

The Spreadsheet That Told the Truth

At HealthFusion, in the early days, we had:

• Four senior engineers

• An ANT script

• And a lot of copy-paste

Deployments rotated like a chore wheel—we rotated amongst us: move files, run scripts, verify manually, pray nothing slipped.

Because I'm me, I tracked it in a spreadsheet.

What it showed was plain: by the time you added up all that deployment babysitting and manual verification, we were already silently paying for a DevOps engineer—just fragmented across people who were hired to do something else.

We hadn't optimized for outcomes. We had optimized for "we're busy, so it must be fine."

That's when I went to my then-boss with the numbers and said: we're on VSS, budgets are tight, constraints everywhere—but where there is will, there is a way.

We need a DevOps engineer.

The Investment That Changed Everything

Enter Hue Nguyen.

Hue fit into our constraints without drama. He didn't ask for perfect conditions; he helped us build better ones. That's what good engineers do: they don't bend to constraints, they work through them.

Around the same time, we were deep in the GitFlow vs trunk debate. I've never believed in over-parenting engineers. Trunk felt riskier, but honest. Less ceremony. More ownership.

We chose trunk-based development.
We moved to Bitbucket.
And over the next ten months, we stitched together real CI/CD.

To gently brainwash our engineers (lovingly!) into embracing CI/CD, TDD, and trunk-based development, we even had a Deming's Principle slide deck looping on the engineering TV.

The slides flashed quotes like "Bring the pain forward" from Jez Humble and "Bake quality in" from Deming.

My favorite was a cartoon of a chicken on an assembly line, looking confused at a carton instead of an egg—a reminder that testing isn't something you bolt on later.

We laughed about it every day, but somewhere between the laughter and the repetition, the mindset clicked.

"Bring the pain forward" became our quiet battle cry for speed with discipline.

Those were fun days—a little chaotic, a little revolutionary.

DORA Before DORA: What We Actually Measured

Years before the DORA metrics became the gold standard for engineering excellence, we were already tracking what mattered:

• Deployment frequency: Weekly releases, every week, without exception

• Lead time for changes: From commit to production in days, not months

• Mean time to recovery: When something broke, we fixed it fast—because our systems were designed for observability and rollback

• Change failure rate: We caught issues before customers did, because we had real QA and real testing infrastructure

This wasn't accidental. It was the direct result of investing in Hue (DevOps), and what would soon follow: investing in other critical disciplines.

We went from quarterly, buggy releases to weekly releases with fast recovery.

Same company. Same product. The difference was: we stopped pretending part-time DevOps was a strategy.

Those numbers didn't just make engineering happy; they de-risked the acquisition. That's the part most teams still underestimate.

When acquisition conversations started, our velocity wasn't just impressive—it was proof we wouldn't slow down post-acquisition. We could show that our platform was maintainable, scalable, and operationally mature.

That matters when someone is writing a $165M check.

This wasn't just about making engineers' lives easier. It was about building a competitive moat. While our competitors were stuck explaining why their latest release was delayed again, we were showing prospects that we shipped improvements every single week.

Years later, as I was preparing to join AMN in late 2021, Dave Baker recommended the book Accelerate during one of our early conversations about how to meaningfully measure SDLC KPIs. I devoured it immediately. For the first time, the research gave language and structure to what we had lived intuitively at HealthFusion. It explained why our practices worked and, in hindsight, revealed that we had accidentally run a textbook Accelerate experiment with the exact outcomes the authors describe. Since then, DORA has been my north star — a clear, research-backed compass for measuring team velocity, guiding engineering strategy, and helping executive leadership teams orient around predictable delivery. Every time I’ve anchored organizations around DORA, I’ve seen the same pattern repeat: clarity rises, firefighting drops, and outcomes improve.

The Power of Real Guardians

In parallel, I invested in people and tools that behaved like Lego blocks for quality and reliability.

Jim Stephens joined us after two decades at Intuit. A calm, seasoned presence when many would've chosen retirement. I'm still grateful he said yes. He brought QA discipline that transformed how we thought about quality—not as a gate, but as a continuous practice woven into development.

John McDonald, our DBA, changed the physics of what we could even consider.

Before him: We were Google-powered DBAs, guessing, patching, hoping.

With him: We could ask real questions. We could optimize queries with confidence. We could design for scale without praying our way through migrations.

He gave structure where there was only courage and duct tape.

And thanks to Yuri and Gabe, who came up with the most creative solutions to build an A/B environment on impossibly tight budgets. Without Yuri’s infrastructure instincts and Gabe’s willingness to experiment, the dream of CI/CD would have stalled. What followed was a true symphony — engineering, DevOps, QA, infrastructure, and application support working together across disciplines to get us to the finish line. It remains one of the best examples of cross-functional craft I’ve seen.

These people did not stand on keynote stages. They just made sure everyone else could.

When I rewind the HealthFusion journey, I don't only see architecture and vision. I see Hue, Jim, John McDonald, Yuri and their team members—the people who made sure everyone else could ship.

What This Investment Actually Bought Us

Let me be concrete about the ROI:

Time to market: Feature requests to production in weeks, not quarters

Customer trust: Predictable release cadence meant customers could plan around our updates

Engineering retention: Our best engineers stayed because they could focus on craft, not firefighting

Acquisition confidence: Due diligence revealed operational maturity that justified premium valuation

Post-acquisition success: We didn't slow down after the acquisition because our systems were already built to scale

That decision to invest in the "plumbing"—pipelines, ownership, DevOps, QA, database excellence—was an inflection point.

It's one of the reasons we could demonstrate operational maturity and predictable velocity during due diligence.

It's one of the reasons our story ended in a $165M acquisition.

ForeSee: When You Start From the Right Instinct

At ForeSee Medical, I arrived already converted. Greenfield helped. It was natural to say from day one:

We will test.
We will automate.
We will wire CI/CD into how we work, not bolt it on later.

Scott Schreckengaust wrote the tox script and internal tooling that kept us honest.

Shruti Bhandari, the lone QA champion—truly a one-woman army—caught what others would've waved through.

Ernesto Ruy Sanchez Valenzuela pulled me fully into Infrastructure as Code; his attention to automation detail likely made our compliance journey far more sane than it could have been.

And yes, we brought back the TV again. This time, it displayed the PR queue in real time. No fancy dashboards, no Slack nags—just pure visibility. When everyone could see which pull requests were waiting for review, no one wanted to be that person holding up the flow. Review times dropped, quality rose, and culture quietly shifted.

Again, the same pattern: quiet technical guardians, loud compounding effects.

We were building an AI-powered HCC Risk Adjustment platform from scratch using NLP and Deep Learning. The technical complexity was immense. But because we invested in these disciplines from the start, we could move fast and maintain quality.

Because those disciplines were present from day zero, we never had to 'slow down to go fix the foundations.' The speed was built-in, not bolted on.

MDLand: The Pattern Surfaces Again

At MDLand, I saw both sides of the pattern.

When Mujeeb Mohammed (Mike) came in as DBA and partner:

• 20-year-old databases moved

• Complex estates landed cleanly on modern AWS SQL Server and MongoDB

• Edge cases were treated with care, not guesswork

It didn't feel like a bolt-on vendor. It felt like acknowledging reality: "This is hard and important. It deserves a real owner."

With Mike and the team, we moved the entire data center to AWS—migrating decades of production data without drama because we had the right specialists doing the work they were trained to do.

Once again, I was reminded: these roles are not optional upgrades. They are structural.

The Anti-Pattern: When Everyone Is Everything (And No One Really Is)

I've also seen the opposite pattern play out across different stages and contexts.

Engineers "also doing":

• DevOps

• QA

• DBA

• Security

The old "Google-powered DBA" has become the "ChatGPT-ed DBA."

Different tools, same fragility.

Let me be clear: I am not criticizing engineers.

If anything, I have profound respect for the way they keep saying "yes":

"I'll handle the deploy."
"I'll own the release."
"I'll debug prod at midnight."
"I'll look at the indexes."
"I'll skim the security checklist too."

But here's the uncomfortable truth:

When brilliant engineers are forced into doing four jobs—out of habit, fear, or insecurity about "needing help"—you don't get four excellences; you get four half-finished disciplines.

The system is weaker.
The releases are riskier.
The real craft of DevOps, QA, DBA, security never takes root.
And the engineers themselves don't get to fully shine where they're strongest.

What This Actually Costs

Let me tell you what happens when organizations don't invest in these disciplines:

You hire more developers and ship slower. Your best engineers become permanent firefighters. Security turns into theatre. Audits are panic, not confirmation. And your strongest people quietly leave for teams that take specialization seriously.

I've seen this play out. The cost isn't just technical debt—it's enterprise value walking out the door.

A Gentle Nudge to Both Sides

This is where I want to gently nudge both engineers and leaders.

To Engineers:

It is not a demotion to say, "We need a real DBA."

It is not a weakness to say, "We need someone to own pipelines properly."

It is not an ego loss to say, "I don't want to be security theatre; I want a partner who lives and breathes it."

Be a little selfish.

You were not hired to be five departments in one body.

When you advocate for specialized support, you're not admitting failure—you're advocating for structural excellence. You're making space for yourself to do the work you were actually hired to do, at the level of craft you're capable of.

And organizations that listen? They ship faster, with better quality, and build more value.

To Leaders:

If your culture makes your best engineers afraid to ask for that help, you have not built a lean team—you've built a ticking time bomb.

Look at your last three production incidents.
Look at your deployment timeline.
Look at your unplanned work backlog.

Ask yourself: are we asking engineers to be heroes, or have we built a system where excellence is structurally supported?

The answer will tell you where to invest next.

Stop confusing heroic multitasking with a healthy engineering system.

Look at where your operational backbone is being bludgeoned by "we'll just have the engineers cover it."

Budget for the specialized disciplines before you build another glass wall.

Your future self—and your innovation engine—already know this is the right answer.

A Sacred Duty (And Deep Gratitude)

So in this pause, during this sabbatical, I'm connecting the dots across all these chapters:

The DevOps engineers.
The DBAs.
The QA champions.
The infrastructure specialists.

The humans wiring automation, watching logs, reading the fine print on compliance, carrying risk on their shoulders so others can move fast.

They are not budget line items.
They are not a "nice to have once we're bigger."
They are your lifeline of innovation.

To the guardians of the galaxy I've worked with—

Hue, Jim, John McDonald, Scott, Shruti, Ernesto, Mike—and the many others whose names I haven't written here but whose work I carry with me—

Thank you for quietly holding up the world so others could ship, pivot, demo, and dream.

Your work didn't just enable features. It enabled velocity. It enabled trust. It enabled enterprise value.

You are the reason a $165M exit was possible.
You are the reason we could move from quarterly chaos to weekly excellence.
You are the reason innovation became sustainable instead of heroic.

Acknowledgment

Gratitude to my then bosses at HealthFusion (Seth, Sol, Jon) who trusted the analysis and supported the hires even when budgets were thin—the outcomes spoke for themselves.

And to every engineer, QA professional, DevOps specialist, DBA, and product partner who came along for the journey across HealthFusion, ForeSee Medical, and MDLand—thank you for showing up for the craft, for the laughter, and for the learning.

These lived experiences shaped how I now understand innovation—not as heroics, but as the quiet discipline of building strong foundations together.

The Bottom Line

We built the house. It looked perfect—glass walls, open layout, solar panels.

Then the first rain came… and nothing drained.

Don't build glass walls before you invest in plumbing.

Don't celebrate your product roadmap while ignoring the operational backbone that makes shipping possible.

Invest in the guardians. Build the innovation engine. Create the enterprise value.

The DevOps engineers, the DBAs, the QA champions, the infrastructure specialists—they're not overhead.

They're the foundation of everything you're trying to build.

And if my experience has taught me anything, it's this:

When you invest in foundations, you don't just build faster. You build something resilient enough that, when growth or acquisition conversations happen, your systems strengthen the story instead of undermining it.

Rumpa Giri is a healthcare technology executive with 20+ years of experience building and scaling engineering teams. She's led technology at companies acquired for $165M, co-founded AI-powered healthcare platforms, and is currently exploring the intersection of agentic AI and healthcare operations during a strategic sabbatical.