Ravindra BagaleCourses & study guides मराठी Track your progress

Guides

DevOps vs SRE vs Platform Engineer: Which Role Fits You?

Job posts mix DevOps, SRE and Platform Engineer. Same clouds, different daily focus. This guide compares duties, tools and interview questions so you pick a fit — with stories from Infosys-style accounts and Google/Amazon-style reliability teams.

Friends, titles change every year. Ask what you will own on call. Chala mitrano — pick the pain you like to fix.

Quick answer

Hold this vertical picture:

  1. DevOps owns shipping pipelines.
  2. SRE owns reliability and toil.
  3. Platform owns the paved road for many teams.
  4. Match your pager pain before you apply.
  5. Practice the interview style for that title.

Language: English · मराठी · हिंदी

DevOps EngineerShip: CI/CD, build, deploy SREReliability: SLOs, toil, incidents Platform EngineerShared tools for many teams Your fitMatch skills + pager pain
Three roles — different daily focus

Why titles confuse freshers

At Infosys, TCS, Wipro, Amazon, Google and product startups, the same tooling shows up under three names. A Pune fresher applies to "DevOps" and the JD talks about SLOs. A Bengaluru candidate sees "Platform" and the interview asks about Jenkins.

Why: Titles are marketing; daily work is ownership. What: Three roles with different primary pain. How: Read the duties below, then match your story.

Concept 1: DevOps Engineer — ship the change

Why: Business needs features in production without weekend panic. What: CI/CD, Docker/K8s deploy, infra as code, basic monitoring. How: Own a pipeline from commit to a known environment.

Example 1: Check a pipeline job status locally (stand-in)

Type this:

printf '%s\n' 'stage=test' 'result=SUCCESS' > /tmp/ci-status.txt
cat /tmp/ci-status.txt

You will see:

stage=test
result=SUCCESS

What it means: A green stage means tests ran. Real Jenkins/Actions show the same idea in a UI.

Example 2: Tag a release the way shipping teams do

Type this:

cd /tmp && mkdir -p ship-lab && cd ship-lab
printf 'v1\n' > VERSION
git init -q -b main
git add VERSION
git -c user.email=a@b.c -c user.name=Lab commit -q -m "release: 1.0.0"
git tag -a v1.0.0 -m "ship 1.0.0"
git tag -l

You will see:

v1.0.0

What it means: DevOps often owns release tags and deploy jobs.

Example 3: Confirm the artifact exists

Type this:

test -f /tmp/ship-lab/VERSION && echo ARTIFACT_OK

You will see:

ARTIFACT_OK

What it means: Shipping means a known build leaves the laptop.

Concept 2: SRE — keep the promise

Why: Users at Netflix-scale or Flipkart sale nights care about errors and latency, not deploy logos. What: SLIs/SLOs, error budgets, toil reduction, incident response. How: Define one SLO, measure it, page only when the budget burns.

Example 4: Simple availability math

Type this:

python3 -c "ok=995; total=1000; print(round(100*ok/total,2))"

You will see:

99.5

What it means: If SLO is 99.9 percent, 99.5 percent burns budget.

Example 5: Error budget left (toy)

Type this:

python3 -c "slo=99.9; actual=99.5; budget=100-slo; used=slo-actual; print('budget_pct_left', round(100*(budget-used)/budget,1))"

You will see:

budget_pct_left 60.0

What it means: SRE talks in budget left, not gut feel.

Example 6: Toil signal — same ticket again

Type this:

printf '%s\n' 'restart nginx' 'restart nginx' 'restart nginx' | sort | uniq -c

You will see:

3 restart nginx

What it means: Three identical ops = toil. Automate or document once.

Concept 3: Platform Engineer — shared path

Why: Twenty product teams should not invent twenty Jenkinsfiles. What: Internal developer platforms: templates, golden paths, self-service. How: Provide one paved road: repo template + pipeline + environments.

Example 7: Count teams on one template

Type this:

printf '%s\n' team-a team-b team-c > /tmp/platform-tenants.txt
wc -l < /tmp/platform-tenants.txt

You will see:

3

What it means: Platform success = many teams on one path.

Example 8: Golden path checklist

Type this:

printf '%s\n' 'repo-template=yes' 'ci-template=yes' 'dev-env=yes' 'prod-gate=yes' > /tmp/golden.txt
cat /tmp/golden.txt

You will see:

repo-template=yes
ci-template=yes
dev-env=yes
prod-gate=yes

What it means: A paved road has defaults, not a wiki dump.

Example 9: Self-service ticket avoided

Type this:

python3 -c "manual=12; after=3; print('tickets_saved', manual-after)"

You will see:

tickets_saved 9

What it means: Platform work is measured in tickets removed.

Concept 4: Pick your fit with evidence

Why: Wrong title wastes six months of prep. What: Match skills + pager pain + interview focus. How: Write three bullets: ship story, reliability story, platform story. Keep the strongest.

Example 10: Skill map file

Type this:

printf '%s\n' 'linux=strong' 'ci=medium' 'slo=weak' 'k8s=medium' > /tmp/skill-map.txt
grep strong /tmp/skill-map.txt

You will see:

linux=strong

What it means: Start where you are strong; grow the weak line with a lab.

Example 11: Interview focus reminder

Type this:

printf '%s\n' 'DevOps: pipeline debug' 'SRE: incident + SLO' 'Platform: template design' > /tmp/focus.txt
cat /tmp/focus.txt

You will see:

DevOps: pipeline debug
SRE: incident + SLO
Platform: template design

What it means: Practice the question style for the title on the JD.

Example 12: One-sentence pitch

Type this:

printf '%s\n' 'I own CI from commit to ECS and cut failed deploys by half in a lab.'

You will see:

I own CI from commit to ECS and cut failed deploys by half in a lab.

What it means: Concrete ownership beats buzzwords.

Side-by-side (remember this)

Role Primary pain Typical tools Interview asks
DevOps Slow or risky releases Jenkins/Actions, Docker, Terraform Pipeline failure debug
SRE Outages and toil Prometheus, SLOs, runbooks Incident timeline
Platform Many teams reinvent wheels Backstage-style portals, templates Golden path design

Suppose we are…

Infosys / TCS account

Client title says DevOps. Work may still be Jenkins + Linux. Ask about on-call.

Google-style SRE

Expect error budgets and postmortems.

Startup "Platform"

You may still write Terraform — but for other teams.

Ravindra Bagale's Tip

In the interview at Google-style SRE or Amazon DevOps, say which pager pain you enjoy. Reliability, shipping, or shared platforms — pick one story.

Check yourself (quiz)

Q1. What does a DevOps engineer mainly own?

Shipping: CI/CD, builds and deploys to known environments.

Q2. What number does an SRE watch closely?

SLOs and error budget left.

Q3. What is a golden path?

A default template for repo, CI and environments that many teams share.

Q4. Why ask about on-call on an Infosys-style project?

The title may say DevOps while the pain is night restarts.

Q5. How do you pick a fit?

Match skills plus the pager pain you can talk about with a real lab story.

Reading tracking

Mark this guide when you finish. Your browser remembers it on this device.

Titles differ by ownership. DevOps ships. SRE protects the promise. Platform paves the road. Choose with evidence, not logos. Did you understand? Let us try it once more.

Frequently asked questions

DevOps vs SRE vs Platform — short answer?

DevOps owns shipping pipelines; SRE owns SLOs and toil; Platform owns shared golden paths for many teams.

Do they use different clouds?

Often the same tools; the daily ownership differs.

What should Infosys-style candidates ask?

What on-call looks like — the title may say DevOps while the pain is night restarts.

What is a golden path?

A default template for repo, CI and environments that many teams share.

How do I pick a fit?

Match skills plus the pager pain you can prove with a lab story.

Where to read interview questions?

Chapter 70 and the DevOps Interview Hub.