Search for a command to run...
The three phases every engineering team goes through. And why the ratio matters more than the labels.
No comments yet. Be the first to comment.
I've been running DevOps and SRE for five years. QA joined my command in Q3 2025.
I didn't ask for QA. My boss, Chief Product Officer Dinesh, asked me to take it over. So now it's three pillars under one roof — QA, DevOps, SRE — and a team of 19 people operating a platform called Helicarrier.
Helicarrier is Kissflow's internal developer platform on GCP. A unified hub that hosts, governs, and standardizes a fast-growing fleet of internal engineering apps with shared auth, SDKs, and automated operations. As AI coding tools let more teams build apps, Helicarrier is where those apps come to life.
I don't lead from the sidelines. I'm in the postmortems. I'm in the war rooms. I'm in the 2 AM incident calls. Our evolution isn't theoretical — it's built on real operational experience and engineering depth.
And I don't lead alone.
Every pillar has a right hand and a left hand — leaders who drive execution while I drive vision. In DevOps and SRE, that's Noordeen and Aravindhan. In QA, it's Santhosh and Akshaya. They run the floor. I run the direction.
But this story isn't about the floor. It's about what happens when you stop running in place — and start running toward something new. ⚡
If you've led engineering teams long enough, you've seen this pattern. Maybe you're in it right now. Every team goes through three phases. Ours did. And here's the thing most people miss — these phases don't replace each other. They coexist.
But let me start at the beginning.
The era of survival.
Every team starts here. You're keeping systems alive, releases moving, fires contained. Every person running hard just to stay in place.
QA lived in manual test execution. Regression suites run by hand. Bug reports written in free-form text with no structure. No measurement of investigation quality — only ticket count.
The engineer who spent three days on a deep root cause analysis got the same recognition as someone who closed ten simple tickets. Effort was invisible.
DevOps ran on deployment scripts executed manually. Environment provisioning took days. Every release was a controlled emergency. Toil was the default operating mode — and nobody had time to engineer their way out of it.
SRE lived in three dimensions — Prevention, Detection, Remediation. In the OE era, we were almost entirely stuck in remediation. Our in-house pager firing at 2 AM. Post-mortems written after the damage. Prevention was aspirational. Detection was basic.
We were always responding. Never preventing.
The hidden cost? Burnout. Frustration. People doing incredible work in silence — fighting fires nobody saw, writing root cause analyses nobody read, deploying fixes at midnight.
The system was blind to quality. It only counted speed.
I won't romanticize it. OE is necessary. You can't skip it. But if you stay in it too long, you lose your best people to exhaustion — and you never build the muscle for what comes next.
The shift from reactive to proactive.
We stopped just running — and started building the systems that run themselves.
QA moved from gatekeeping to engineering. Regression suites automated. Quality gates in CI. Coverage measured. Test frameworks built. QA stopped being the last checkpoint before a release and started being an engineering discipline.
DevOps moved from toil to pipelines. Jenkins orchestration. Infrastructure as code. Automated deployments and provisioning in minutes, not days. Toil systematically eliminated across the delivery lifecycle.
SRE moved from remediation to all three dimensions. SLOs defined. Structured monitoring strengthened detection. Proactive post-mortems improved prevention. All three dimensions — prevention, detection, remediation — now engineered, not improvised.
The result? Systems improved. Reliability improved. Quality improved.
But humans were still the bottleneck.
Every decision still flowed through a human. Every insight required someone to look at the data and think. Every analysis was written by hand. Every pattern spotted by instinct.
We'd built good systems. Reliable systems.
But I needed systems that could think for themselves.
This is where we're heading now.
Vibe Engineering is the era of intelligence. AI-driven workflows across all three pillars. Systems that don't just execute — they understand, adapt, and decide.
QA becomes intelligent testing. AI scores investigation depth. Detects resolution patterns. Evaluates quality at scale. QA becomes a quality intelligence layer — not a gate, not a checkpoint, but a system that sees what humans miss.
DevOps becomes autonomous pipelines. Self-healing infrastructure. Predictive scaling. Intelligent deployment decisions. The engineer becomes the orchestrator, not the operator.
SRE becomes predictive reliability. Prevention, detection, remediation — all AI-augmented. Causality chains for prevention. Anomaly detection before impact. Automated remediation at speed. SRE shifts from responding to anticipating.
Vibe Engineering doesn't replace the engineer. It amplifies the engineer into an orchestrator.
But here's the nuance that most people get wrong about transformation narratives. 👇
OE, EE, and VE — Operational Efficiency, Engineering Excellence, and Vibe Engineering — aren't a linear replacement. They always coexist.
In the life of DevOps, SRE, and QA, it will always be a blend. The ratio shifts — but all three live together.
In the past: OE was 75% of our effort. EE was 20%. VE was barely 5%. Firefighting dominated. No bandwidth for anything else.
Now: OE is 30%. EE is 45%. VE is 25%. Engineering excellence has taken hold. Vibe Engineering is emerging.
In the future: OE will be 10%. EE will be 30%. VE will be 60%. Intelligence leads.
Here's the insight that changed how I think about this: VE accelerates the entire transformation. It doesn't wait for perfection. It enables it.
In the past, firefighting didn't leave room for excellence. You can't build automation when you're busy putting out fires. But with VE emerging, we can swiftly shift from OE to EE, then from OE and EE into VE. Intelligence reduces toil. Toil reduction creates space for engineering. Engineering feeds back into better intelligence.
It's a flywheel. And once it starts spinning, the ratio shifts fast.
I had the vision. I had the technology — Digital Intelligent Agents capable of reasoning, scoring, deciding.
What I didn't have was proof.
Proof that a small team — cross-functional, AI-augmented, working at speed — could build something that changes how we see our own people.
Not a theoretical framework. Not a slide deck. Not a strategy document.
A working system. Built from scratch. With real data. In real time.
I needed a squad. ⚡
That's what Part 2 is about. How I picked two engineers — one from QA, one from DevOps — and formed the Viberanium Squad. Their first warm-up mission. And why timing and direction matter more than individual talent.
👉 Part 2: The Assembly — dropping Monday, April 7.
📌 This is Part 1 of 7 in the Viberanium Squad series.
Swami K — AVP, DevOps, SRE & QA at Kissflow
Build Fast. Build Real. Ship It. ⚡
