Skip to the notes
JSGroundwork
JSGroundwork handwritten · web dev
✎Playground→⌘Problems↻Review🔥Progress

Chapters

20 chapters
⌕
Beginner15›
00Scouting reportR1Screening callR2Machine codingR3JavaScript & TSR3·TSTypeScriptR4React & Next.jsR5Node & NestJSR6Databases & RedisR7DSA roundR8System designR9AWS, Docker, CI/CDR10Resume grillingR11BehaviouralR12HR & the number✓The week before
Advanced5›
50LWhat changes at ₹50L50LHard DSA50LDistributed systems50LRuntime internals50LStaff behavioural
/ search[ ] chaptert top

Ready to read

JSJavaScript⑂Git◎Interview prepΣDSA in JSSDSystem Design
More topics15›
</>HTML{ }CSS⚛ReactNNext.jsNeNest.jsTSTypeScriptNoNode.js🐳DockerDBSQL & Databases✓Testing🔒Web Security☁Cloud & DevOps◈GraphQL◆Redis☸Kubernetes
100%
50L

What actually changes at the ₹50L bar

TrackStaff / SDE-3
Timeline2–3 years, two jumps
Real blockerThe resume, not the interview
read before you prep

The ₹20–28L rounds test whether you can build. These test whether you can be trusted with a system nobody fully understands, and with decisions that cost money if you are wrong. Different bar, different preparation, different timeline — and one thing you already have that most candidates do not.

S0.1
The five gaps between where you are and ₹50L
What they are really testingHonest calibration. Prepping the wrong things for a year is the expensive mistake here.
GapWhere you areWhere the bar is
Scale5,000 users, one region, one databaseMillions of users, sharded data, multi-region, real QPS numbers you have personally watched
DSAPatterns, mediumsTwo rounds, at least one hard. Graphs, advanced DP, heaps, tries — solved live in 35 minutes
DesignCaching, concurrency, a good APIConsistency models, consensus, partition tolerance, exactly-once semantics, capacity maths
BrandThree client studiosRecognisable product company on the resume — unfair, and it is what gets you shortlisted
Scope of influenceYou own four applicationsYou changed how several teams work, and can evidence it

Four of those five are closeable with study. Brand is not — and it is the one that decides whether anybody reads your resume. That is why the path below is two jumps and not one.

S0.2
The two-jump plan, with dates
What they are really testingWhether you optimise for the next offer or for the one after it.

₹12L → ₹50L is not a job change. It is a trajectory. Trying to do it in one move is the single most common way ambitious engineers waste two years applying to roles that never call back.

Now → 3 months

Land ₹22–26L at a real product company. Optimise for scale exposure and a name, not for the top of the band. A ₹24L offer at a company with real traffic beats ₹27L at another studio, and it is not close.

Months 3–24

Get on-call. Own a service with real QPS. Drive one migration that touches another team. Learn the distributed side on the job, where it is free, rather than from a book. Meanwhile: DSA, 4 hours a week, every week.

Months 24–30

Interview at the ₹45–55L tier with two years of product-company scale on the resume, a brand recruiters recognise, and DSA that is warm rather than cold. This jump is normal at that point.

The uncomfortable part: the highest-leverage decision you make in the next three months is which company, not which number. A ₹2L difference today is worth nothing against the ₹25L difference that the right brand and the right traffic make in two years.

2026 note

The one shortcut worth taking seriously: US-remote roles hiring in India. They skip the brand filter, pay well above local bands, and weigh portfolio and communication far more heavily — both of which you are strong on. It is a harder search with fewer openings and it needs referrals, but for your specific profile it is the most realistic way to compress two jumps into one.

S0.3
Your resume is the blocker, not your ability.
What they are really testingNothing. This is the thing to fix first, because everything else is downstream of getting shortlisted.

At ₹22–28L a good recruiter reads your bullets. At ₹50L a screener spends eight seconds looking for a company name they recognise and a scale number that is large. You have neither, so the technical preparation never gets used.

Three things that change that, in order of impact:

  1. Referrals. Above ₹40L, cold applications convert at close to nothing. A referral skips the screener entirely. This is not a networking platitude — it is the actual mechanism, and it is the highest-return hour you can spend.
  2. Lead with Groundwork. You built a learning platform with 85 chapters, 299 runnable exercises, and published tracks in JavaScript, DSA and System Design. Most six-year engineers have nothing like it. Those are exactly the three subjects a ₹50L loop tests — and you are teaching them. Put it in the top third of your resume with the numbers, not in a "personal projects" footer.
  3. Replace vanity numbers with hard ones. "5,000 registered users" is a small number stated plainly. If you have p95 latency, requests per second, data volume, uptime, or cost figures, those read as an engineer who operates systems. If you do not have them, that is a thing to start collecting at your next job from week one.
Say it like this

Alongside my work I write and maintain Groundwork, a technical learning platform — 85 chapters and 299 runnable exercises across JavaScript engine internals, algorithms and system design, with about 15 hours of written material. Teaching a thing at that depth is how I know I actually understand it.

S0.4
What to build in the next six months to close the scale gap
What they are really testingWhether you can manufacture the evidence you are missing, honestly.

You cannot fake having handled millions of users. You can build something where the interesting problems are the ones that only appear at scale, and then talk about them with real numbers. Three that would genuinely move your interviews:

  • Load-test something you own and write up what broke. Point k6 or Artillery at one of your services, find the knee of the curve, fix the first bottleneck, measure again. "I took it from 400 to 3,000 requests per second and here is what the profile showed at each step" is a scale story with real numbers, and it takes a weekend.
  • Build one thing that is genuinely distributed. A job queue with at-least-once delivery, idempotent consumers and a dead letter queue. Small, but it forces you through exactly-once semantics, retries and partial failure — the vocabulary of the design round.
  • Instrument something properly. OpenTelemetry traces, p50/p95/p99 dashboards, an alert on an error budget. This closes attack 4 from the scouting report, and observability is a real interview topic at this level rather than a footnote.

Each of these is small enough to finish and specific enough to talk about for ten minutes. That combination is what makes a side project useful in an interview; a half-built clone of a famous app is not.

←previousHR & the number↑ CovernextHard DSA→