Skip to main content

Chapter 17: Company Playbooks & The First 90 Days

You have the skills. You have the offer — maybe two. Now: which company do you choose, and how do you make sure you survive the first 90 days?

Most engineers treat this decision like a coin toss. They pick the highest number, show up on day one, and hope for the best. Then they wonder why they are PIP'd out in six months or stuck at the same level for three years.

This chapter is your playbook. Not theory. Not "networking tips." A field manual for the three most dangerous phases of your career: choosing where to land, surviving the probation period, and knowing when to leave.

The Company Tier System​

Indian tech companies are not a flat list. They are a tiered ecosystem, and each tier plays by different rules. If you apply Tier 3 strategy to a Tier 1 company, you will fail. If you expect Tier 1 stability at a Tier 3 startup, you will be miserable. Know the game before you sit at the table.

Tier 1: The Global Giants​

Google, Meta, Uber, Stripe, Atlassian, Amazon, Microsoft, Apple.

The Pay. Base salaries for Staff Engineer (L6/E6/IC5 equivalent) range from ₹80 lakh to ₹1.2 crore. Add RSUs worth $150K-$250K per year, and total compensation lands between ₹1.5 crore and ₹2.5 crore. Sign-on bonuses of $30K-$50K are standard. Relocation packages exist but are shrinking post-2023.

The catch: RSUs are dollar-denominated. A 10% rupee depreciation adds ₹8-12 lakh to your annual take-home without you lifting a finger. The reverse is also true — but historically, the rupee only moves one direction.

The Culture. Process-heavy. You will spend 30-40% of your time on design docs, RFCs, and alignment meetings. Code is the output, not the job. At Google, a Staff Engineer's calendar has 15-20 hours of meetings per week. At Meta, you ship fast but review faster — every diff goes through at least two reviewers, and post-incident reviews are brutal and public.

The unspoken rule: your impact is measured by the teams you unblock, not the lines you write.

The Interview Style. Tier 1 companies test fundamentals, not frameworks. Google does 4-5 rounds of DSA + system design + behavioral. Meta adds a "product architecture" round. Stripe has a notorious "bug squash" round where you debug a real codebase live. Amazon's Bar Raiser is a dedicated interviewer whose sole job is to veto candidates who do not raise the bar.

What they all share: they are testing for trajectory, not current knowledge. They want to know if you will be a Staff Engineer in two years, not if you can recite the Node.js event loop today.

The Reality Check. Tier 1 is not for everyone. The bureaucracy is real. Promotions at Google from L5 to L6 can take 3-4 years. You will work with brilliant people, but you will also sit through meetings that could have been emails that could have been a single Slack message. If you need autonomy and speed, Tier 1 will suffocate you.

But the brand pays dividends forever. Two years at Google India opens doors that ten years at a services company never will.

Tier 2: The Indian Unicorns​

Flipkart, Swiggy, CRED, Razorpay, PhonePe, Zerodha, Meesho, Groww, Pine Labs, BrowserStack, Postman.

The Pay. Staff-equivalent roles (Principal Engineer, Architect, Staff Engineer) pay ₹60 lakh to ₹1.2 crore base. Equity is the wildcard. ESOPs at a pre-IPO company like Razorpay or Meesho can be worth ₹2-5 crore at exit — or zero. Post-IPO companies like Zomato and Paytm offer RSUs that trade on the NSE, but the volatility is real. Zomato's stock went from ₹70 to ₹280 and back to ₹120 in eighteen months.

The math you need to do: equity value = (strike price × number of options) vs. (expected FMV at exit × number of options). If the company says "your ESOPs are worth ₹1 crore," ask: at what valuation? At what liquidity event? When? If they cannot answer all three, the number is fiction.

The Culture. Indian unicorns move faster than Tier 1 but slower than startups. You will own more surface area. A Staff Engineer at Swiggy might own the entire order-fulfillment pipeline — something that would be split across three teams at Google. This is a feature, not a bug. You learn faster when you own more.

The downside: Indian unicorns run lean. Headcount is tight. You will be asked to do the work of 1.5 people. Burnout is the number one reason engineers leave Tier 2 companies within two years.

The Interview Style. Tier 2 companies are pragmatic. Flipkart does 3-4 rounds: DSA (medium-hard), system design (real-world, not theoretical), and a "project deep-dive" where you walk through something you built. Swiggy and CRED emphasize system design heavily — they want to know if you can design a food-delivery dispatch system or a credit-card rewards engine, because you might have to.

Zerodha and BrowserStack are outliers. Zerodha barely interviews — they hire based on open-source contributions and a single conversation with the CTO. BrowserStack runs a take-home assignment followed by a code-review round that is more rigorous than most FAANG onsites.

The Equity Math. This is where most Indian engineers get it wrong. Let us walk through a real scenario.

You join Razorpay at a ₹50,000 crore valuation. You get 5,000 ESOPs at a ₹100 strike price. The company IPOs at a ₹1,00,000 crore valuation. Your options are now worth (₹2,000 FMV - ₹100 strike) × 5,000 = ₹95 lakh. After long-term capital gains tax (20% with indexation), you net roughly ₹76 lakh.

But here is what nobody tells you: the average time from Series D to IPO in India is 5-7 years. You will likely leave before the exit. When you leave, you have 30-90 days to exercise your vested options. If you have 3,000 vested options at a ₹100 strike price, you need ₹3 lakh in cash — just to buy shares you cannot sell. Most engineers let their options lapse. Do not be most engineers. Budget for the exercise cost before you sign.

Tier 3: High-Growth Startups and Remote-First​

Seed to Series C companies. Remote-first roles at US/EU startups. Devtools companies, Web3 (if you have the stomach for it), AI-native startups.

The Pay. Base: ₹40-80 lakh. Equity: 0.25% to 1.5% for early employees. The equity is the entire bet. If the company exits at ₹500 crore, 0.5% is ₹2.5 crore. If it dies — and most do — the equity is worth exactly nothing.

Remote-first US startups pay $80K-$150K base for senior IC roles. That is ₹66 lakh to ₹1.25 crore in base alone, no equity needed. The arbitrage is real: you earn a US salary while living in India. But the tradeoffs are real too: night calls, no Indian labor-law protections, and the constant low-grade anxiety that your role could be "rebalanced" to a lower-cost geography.

The Culture. You will ship more code in six months at a Series A startup than in two years at a Tier 1 company. You will also break more things, own more incidents, and have fewer people to blame. The learning curve is vertical. The support structure is thin.

Remote-first companies add another layer: you must be a written communicator. If you cannot write a clear RFC, run an async design review, or debug a production issue over a Loom video, you will struggle. The engineers who thrive in remote-first environments are the ones who treat writing as a first-class skill.

The Interview Style. Startups are inconsistent. Some run FAANG-lite loops. Others do a single take-home and a culture-fit call. The variance is high, and that is the risk: you cannot prepare for every startup's process the way you can for Google's.

What you can do: ask about their interview process in the first recruiter call. If they cannot describe it clearly, that is a red flag. A startup that does not know how to hire is a startup that does not know how to build.

The Risk/Reward Calculus. Tier 3 is for a specific profile: you have financial runway (6-12 months of expenses saved), you are comfortable with ambiguity, and you want the asymmetric upside that Tier 1 and Tier 2 cannot offer. If you have dependents, a home loan EMI, or a low risk tolerance, Tier 3 is not for you — and there is zero shame in that.

How Each Company Interviews​

You cannot prepare the same way for every company. Each one tests a different muscle. Here is the cheat sheet.

Google India (Bangalore/Hyderabad). 4-5 onsite rounds after a phone screen. Two DSA rounds (medium-hard, graphs and DP are common), one system design (distributed systems focus), one behavioral (Googleyness — they want to see intellectual humility and team-first thinking). Staff+ loops add a second system design round and a "cross-functional" round with a PM or TPM. The hiring committee reviews everything after the loop — your interviewers do not decide. The committee does.

Meta (Bangalore, remote). Phone screen (DSA, 2 questions in 45 minutes), then 4 onsite rounds: two DSA (speed matters — 2 questions per round), one system design (product-focused: "design Instagram stories"), one behavioral (Meta's "cultural fit" is about moving fast, being direct, and owning outcomes). Meta's leveling is aggressive — they will down-level you if your system design round is weak, even if DSA is perfect.

Amazon (Bangalore, Hyderabad, Chennai, Pune). The Bar Raiser is in every loop. 4-5 rounds: DSA (practical, not theoretical — trees, arrays, hashmaps, rarely DP), system design (AWS-flavored — they want to see you use their primitives), and 2-3 behavioral rounds. Amazon's behavioral is the most structured in the industry. They use the STAR method religiously. Every answer must have a Situation, Task, Action, Result. They will dig into the "Action" — what you did, not your team. Prepare 8-10 stories with metrics.

Uber (Bangalore, Hyderabad). 4-5 rounds: DSA (hard — Uber's bar is closer to Google than Amazon), system design (real-time systems: design Uber's dispatch, design a payment gateway), and a "craftsmanship" round where you write production-quality code with tests, error handling, and monitoring. Uber's culture is intense. They value ownership and "acting like an owner" — which, in practice, means you are expected to fix problems you were not asked to fix.

Stripe (Bangalore, remote). The bug-squash round is the differentiator. You get a real codebase with failing tests and production bugs. You have 60 minutes to fix as many as you can. They are testing your debugging process, not your speed. Talk through your approach. Use the debugger. Write tests for your fixes. The system design round is API-heavy — Stripe is an API company, and they want to see you think in terms of developer experience, idempotency, and error semantics.

Atlassian (Bangalore, remote). 4 rounds: DSA (medium, practical), system design (collaboration tools — design a real-time document editor, design a notification system), a "values" round (Atlassian's values are specific: "Play as a team," "Don't #@!% the customer"), and a take-home coding exercise. Atlassian's process is slower and more deliberate than FAANG. The bar is high but the pressure is lower.

Flipkart (Bangalore). 3-4 rounds: DSA (medium-hard, standard Indian-tech style — arrays, trees, graphs, DP), system design (e-commerce: design a flash-sale system, design a recommendation engine), and a "project deep-dive" where you present a project you led. The deep-dive is the most important round. They want to see ownership, technical depth, and the ability to drive a project from idea to production. Bring architecture diagrams. Bring metrics. Bring war stories.

CRED (Bangalore). CRED's process is lean: a coding round (practical, not LeetCode — build a small service), a system design round (fintech: design a rewards engine, design a credit-score pipeline), and a culture round with the CTO or VP Engineering. CRED values "high agency" — they want engineers who find problems and fix them without being told. Your behavioral answers should emphasize initiative.

Razorpay (Bangalore). 3-4 rounds: coding (practical, framework-agnostic), system design (payments: design a payment gateway, handle idempotency, design reconciliation), and a "domain depth" round where they go deep on your area of expertise. If you claim to be a Node.js expert, they will ask about the event loop, the V8 GC, clustering, and stream backpressure. Do not bluff.

Remote-First US Startups. The process varies wildly, but a pattern has emerged: a 30-minute culture screen, a take-home assignment (4-8 hours), a code-review round where you walk through your take-home, and a system design round. Some add a "paid trial" — a 1-2 week contract where you work on real tasks. This is the best signal for both sides. If a startup offers a paid trial, take it. You will learn more about the company in two weeks than in twenty interviews.

The First 90 Days​

You signed the offer. You served your notice period. You walked into the office on day one — or logged into Slack, if you are remote.

Now the real work begins.

The first 90 days are not about proving you are brilliant. They are about proving you are reliable. Brilliance gets you hired. Reliability keeps you employed. The engineers who fail probation are rarely the ones who cannot code. They are the ones who cannot navigate the organization, cannot build trust, and cannot deliver something — anything — that the business values.

Here is the 30-60-90 day plan that works.

Days 1-30: Learn the Terrain​

Your only job in the first month is to understand how the company actually works. Not the org chart. Not the onboarding docs. The real machinery.

Week 1: Ship something trivial. Fix a typo in the README. Add a log line. Close a good-first-issue. This is not about impact — it is about learning the deployment pipeline. How does code go from your laptop to production? Who approves PRs? What breaks? You will learn more from shipping a one-line change than from reading a hundred Confluence pages.

Week 2: Map the dependencies. Every system has hidden dependencies. The payment service that goes down every Tuesday at 3 AM because of a cron job nobody owns. The internal library maintained by one engineer who is about to go on paternity leave. The database that has not been vacuumed since 2022. Find these. Write them down. You do not need to fix them yet — you need to know they exist.

Week 3: Build your "who knows what" map. In every organization, there are five to ten people who actually know how things work. They are not always the managers. They are the engineers who have been there for three years, who get paged at 2 AM, who everyone DMs when something breaks. Find them. Buy them coffee. Ask them: "What is the one thing you wish someone had told you in your first month?" Listen. Do not talk.

Week 4: Identify your first real win. By the end of month one, you should have a list of 3-5 problems you could solve in the next 30 days. They should be small enough to ship alone, visible enough that your manager can report them upward, and real enough that they actually help someone. A flaky test that wastes 20 minutes a day. A deployment script that fails 30% of the time. A monitoring gap that caused a missed incident last quarter. Pick one. Ship it in month two.

The Indian Context. In Indian companies — especially Tier 2 and Tier 3 — the first month is also about navigating hierarchy. Indian workplaces are more hierarchical than their Western counterparts. Your manager expects deference, even if they say they do not. This does not mean you should be a doormat. It means you should be strategic about when and how you push back. In month one, do not push back at all. Observe. Learn who has real power (it is rarely the person with the biggest title) and who is just loud.

Days 31-60: Deliver the First Win​

Month two is about execution. You have identified a problem. Now ship the fix.

The Rule of the First Win. Your first win must satisfy three conditions: it must be visible to your manager and their manager, it must be unambiguously yours, and it must ship within 30 days. If it takes longer, it is not a first win — it is a project. Projects come later.

What a good first win looks like:

  • "I noticed our CI pipeline was taking 22 minutes. I parallelized the test suite and added build caching. It is now 8 minutes. That saves roughly 40 engineer-hours per week."
  • "The payment-reconciliation job was failing silently on duplicate transactions. I added idempotency handling and alerting. We caught ₹12 lakh in duplicate charges in the first week."
  • "Our error logs were unstructured JSON blobs. I added structured logging with request IDs. The on-call team can now trace a request end-to-end in under 30 seconds."

What a bad first win looks like:

  • "I refactored the authentication module to use a cleaner pattern." (Invisible. No metric. Nobody asked for it.)
  • "I wrote a proposal for migrating our entire stack to Kubernetes." (Too big. Will not ship in 30 days. Looks like you are avoiding real work.)
  • "I fixed 15 small bugs across 8 different services." (Scattered. No single narrative. Hard for your manager to sell upward.)

The Trust Equation. Trust = (Credibility + Reliability + Intimacy) / Self-Orientation. This is the Trust Equation from David Maister's The Trusted Advisor, and it applies perfectly to engineering organizations.

Credibility: Do you know what you are talking about? Built through code quality, design reviews, and technical decisions that hold up under scrutiny.

Reliability: Do you do what you say you will do? Built through shipping on time, communicating delays early, and never surprising your manager.

Intimacy: Do people feel safe being honest with you? Built through 1:1s where you listen more than you talk, admitting when you do not know something, and never throwing anyone under the bus.

Self-Orientation: Are you in it for yourself or for the team? This is the denominator. The more you talk about your own career, your own impact, your own promotion — especially in the first 90 days — the lower your trust score. Focus on the team's problems. Your career will follow.

Days 61-90: Expand Your Surface Area​

By month three, you have shipped a win and built baseline trust. Now expand.

Start saying no. The most dangerous word in your first 90 days is "yes." Everyone will ask you for things: code reviews, design feedback, interview loops, onboarding buddies, "can you just take a quick look at this?" If you say yes to everything, you will do nothing well. Start practicing: "I would love to help with that, but I am focused on X this week. Can we revisit next sprint?"

Find a mentor who is not your manager. Your manager has a conflict of interest: they need you to deliver for the team, which is not always the same as what is best for your career. Find a Staff or Principal Engineer who has been at the company for 2+ years. Ask for a 30-minute chat every two weeks. Come with specific questions, not "how do I grow?" Ask: "I noticed the team struggles with X. How did you handle that when you were at my level?" or "What is the one skill that separates the engineers who get promoted here from the ones who do not?"

Understand the political landscape. Every company has politics. "I do not do politics" is not a strategy — it is a decision to be blindsided by politics you did not see coming. You do not need to play the game. You do need to understand it.

Map the power structure: Who controls headcount? Who controls the roadmap? Whose opinion does your skip-level manager trust? Who is rising and who is checked out? You do not need to act on this information. You need to have it, so that when reorgs happen — and they always happen — you are not surprised.

The Story of Rohan. Rohan joined a Tier 2 fintech as a Staff Engineer in 2023. His first month, he noticed the payment-processing pipeline had a race condition that caused duplicate charges roughly 0.03% of the time. The team had known about it for eight months. Nobody had fixed it because it was "too risky" — the code was old, the original author had left, and nobody wanted to touch it.

Rohan did not fix it in month one. He spent month one understanding the pipeline, building relationships with the SRE team, and writing a rollback plan. In month two, he shipped the fix with a feature flag, a canary deployment, and a dashboard tracking duplicate rates. The fix went live. Duplicate charges dropped to zero. His manager's manager mentioned it in the quarterly business review.

Rohan was promoted to Senior Staff in eighteen months. Not because he was the best coder on the team — he was not. Because he did the thing everyone else was afraid to do, and he did it in a way that made his manager look good.

That is the game. Play it or do not. But know that it is being played.

The Growth Trajectory: Staff to Senior Staff to Principal​

You survived the first 90 days. You have been at the company for a year. Now what?

The jump from Staff to Senior Staff is harder than the jump from Senior to Staff. The jump from Senior Staff to Principal is harder still. Each level demands a fundamentally different way of working.

Staff Engineer (L6/E6/IC6). You own a team-level problem. You are the technical anchor for a squad of 5-10 engineers. Your code is still visible — you write critical-path features, review complex PRs, and set technical direction. Your impact is measured in team velocity and system reliability. You spend 40% of your time coding, 30% reviewing and mentoring, 30% in design and alignment.

Senior Staff Engineer (L7/E7/IC7). You own an org-level problem — 3-5 teams, 20-50 engineers. You write almost no code. Your output is design docs, RFCs, technical strategy, and unblocking other Staff Engineers. Your impact is measured in org velocity: did the teams you support ship faster and better because of your work? You spend 10% coding (prototypes, critical fixes), 40% designing and reviewing, 50% in meetings, mentorship, and cross-org alignment.

This is the hardest transition for most engineers. You have to let go of coding as your primary identity. If you cannot — if you need the dopamine hit of merging PRs — you will stall at Staff forever.

Principal Engineer (L8/E8/IC8). You own a company-level problem. You set technical strategy for 50-200 engineers. You identify problems before anyone else sees them. You influence without authority across multiple VPs. Your output is direction: the 2-3 year technical vision, the big bets, the architectural decisions that will either make or break the company's engineering velocity for the next five years.

Principal Engineers are rare. Most companies have fewer than five. The path is not linear — you do not get promoted to Principal by doing more of what got you to Senior Staff. You get there by seeing around corners, by building coalitions, by making bets that pay off.

The Indian Reality. In India, the Staff+ market is still maturing. Tier 1 companies have well-defined IC tracks — you can reach Principal without managing anyone. Tier 2 companies are catching up. Many still conflate "seniority" with "headcount" — they will make you a manager whether you want it or not. If you want to stay on the IC track, ask about this explicitly during interviews. "What does the IC track look like beyond Staff? Who is the most senior IC at the company, and what did they do to get there?" If the answer is vague, the track does not exist.

The Exit Strategy​

Here is a truth nobody puts in the onboarding doc: you will leave every company you join. The only question is when, how, and on whose terms.

When to stay. Stay when you are learning faster than the company is growing. Stay when your manager invests in your career — gives you stretch assignments, advocates for your promotion, tells you hard truths in private. Stay when the equity is real and the liquidity event is within 18-24 months. Stay when you wake up on Monday and do not dread opening Slack.

When to leave. Leave when you have been at the same level for 2+ years with no clear path to promotion. Leave when your manager is a blocker, not an accelerator — they take credit for your work, they cancel your 1:1s, they cannot articulate what "good" looks like at the next level. Leave when the company's best engineers are leaving. (The best signal in tech: track where your most respected colleagues go. If three of them leave in six months, you should be polishing your resume.)

Leave when the equity is underwater — your strike price is higher than the current 409A valuation, and there is no credible path to an up round. Leave when you have stopped learning. The day you stop learning is the day your market value starts decaying.

How to leave. Give notice after your RSUs vest, not before. Do not badmouth the company in your exit interview — it buys you nothing and can burn references. Write a thorough handoff doc. Train your replacement if there is one. The tech industry in India is small. The engineer you throw under the bus today might be the hiring manager at your dream company in three years.

The Two-Year Rule. In the ₹60 lakh to ₹1 crore range, the optimal tenure is 2-3 years per company. Less than 18 months looks like job-hopping. More than 4 years at the same level looks like stagnation. The exception is Tier 1 companies, where 4-5 year tenures are normal and expected — the promo cycles are longer, and the brand value compounds.

The Compounding Effect. Every jump should increase your scope, not just your salary. Staff at Flipkart to Senior Staff at a Series C startup. Senior Staff at a startup to Principal at a Tier 1. Each move should expand the problems you are trusted to solve. If you make three lateral moves in a row — same level, different company, slightly more money — you are not building a career. You are running in place with a nicer treadmill.

The System, Not the Sprint​

You now have the playbook. Company tiers, interview strategies, the 30-60-90 day plan, the growth trajectory, the exit calculus. But a playbook is only as good as the person executing it.

The engineers who reach ₹1 crore do not do it by working harder than everyone else. They do it by working on the right things, at the right companies, with the right strategy. They treat their career like a system, not a series of sprints. They optimize for learning in their 20s, for scope in their early 30s, and for leverage in their late 30s. They know when to go deep and when to move on.

You have the same information they do now.

Remember the engineer who opened this book, wondering if ₹1 crore was even possible? Look in the mirror. That engineer is gone. The person looking back has the roadmap, the skills, and the system. You know which companies pay what. You know how to survive the first 90 days. You know when to stay and when to walk.

Now go execute.