How Tech Interviews Work

The tech interview process is a funnel. Understanding each stage — what it tests, who evaluates you, and how decisions get made — lets you allocate preparation time where it actually matters. Most candidates fail not because they lack skill, but because they misunderstand what a given stage is selecting for.

The Funnel

A typical hiring pipeline at a mid-to-large tech company looks like this:

100 applications received
 40 pass resume screen (ATS + recruiter review)
 15 get recruiter calls
 10 pass phone screen
  3 advance to onsite
  1 receives offer

The exact ratios vary by company, role, and market conditions, but the shape is consistent. Each stage is designed to cheaply eliminate candidates who would fail a later, more expensive stage.

Stage 1: Resume Screen

This is the first filter and the most automated. At large companies, an Applicant Tracking System (ATS) parses your resume for keywords, years of experience, and degree requirements before a human ever sees it.

What gets you through:

Relevant keywords matching the job description
Quantified achievements (numbers, percentages, scale)
Clean formatting that ATS can parse (no tables, no columns, no headers/footers)
Referral flags — a referral moves you past ATS entirely at most companies

What gets you rejected:

No match on core technical requirements
Typos in the first three lines (recruiters scan, they do not read)
Generic resumes not tailored to the role
Gaps with no explanation (addressed in cover letter or LinkedIn)

A recruiter spends 6-10 seconds on your resume. They are looking for signal: company names they recognize, technologies that match, and impact metrics. Everything else is noise.

Stage 2: Recruiter Call

Duration: 15-30 minutes. This is a logistics and calibration call, not a technical assessment. The recruiter wants to confirm:

- You are actually interested in this specific role
- Your experience roughly matches what the resume claims
- Your salary expectations are within the band
- Your timeline works (start date, notice period, competing offers)
- You can communicate clearly in English (or whatever the working language is)

Common mistakes at this stage: being unprepared to discuss your own resume, not having questions about the role, or revealing that you applied to 50 companies and cannot remember which one this is.

The recruiter is also selling. They want you to be excited. Match their energy.

Stage 3: Phone Screen

Duration: 45-60 minutes. Usually one technical interview conducted over video call with a shared coding environment (CoderPad, HackerRank, or similar).

What it tests:

Can you solve a medium-difficulty algorithm problem in 30-40 minutes?
Can you communicate your thought process while coding?
Do you ask clarifying questions before diving in?
Can you test your own code and find bugs?

The bar here is lower than the onsite. The interviewer is looking for "would this person be a reasonable use of the onsite panel's time?" You do not need a perfect solution. A working solution with clear communication beats a perfect solution arrived at in silence.

Typical phone screen structure:
  0-5 min:  Introductions, interviewer describes the problem
  5-10 min: You ask clarifying questions, discuss approach
  10-40 min: You code the solution, talking through decisions
  40-50 min: You test your solution, discuss edge cases
  50-60 min: Your questions for the interviewer

Stage 4: Onsite (Virtual or In-Person)

The onsite is the main event. It typically consists of 4-6 interviews over a half or full day. The exact mix depends on the company and level, but a common structure is:

Interview 1: Coding (data structures & algorithms)
Interview 2: Coding (different problem type)
Interview 3: System design (mid-level and above)
Interview 4: Behavioral / leadership principles
Interview 5: Coding or domain-specific (varies)

Coding Rounds

You will face 1-3 coding rounds. Each is 45-60 minutes with one or two problems. The interviewer evaluates:

Problem-solving approach and ability to break down ambiguous problems
Code quality — readable, well-structured, reasonable variable names
Correctness — does it actually work, including edge cases
Efficiency — can you analyze and improve time/space complexity
Communication — did you explain your thinking throughout

System Design

For engineers with 3+ years of experience, at least one round will be system design. You will be asked to design a system at a high level: "Design Twitter," "Design a URL shortener," "Design a rate limiter."

What it tests: Can you think about systems at scale? Can you make reasonable tradeoffs? Do you understand distributed systems concepts (load balancing, caching, database sharding, message queues)?

Behavioral

Every major tech company has behavioral interviews, though they call them different things. Amazon calls them "Leadership Principles" interviews. Google calls them "Googleyness & Leadership." Meta calls them "Culture fit."

What it tests: How you work with others, handle conflict, deal with failure, prioritize under pressure, and influence without authority. You will be asked to tell stories from your past experience.

Common behavioral question patterns:
  "Tell me about a time when..."
  "Describe a situation where..."
  "Give me an example of..."
  "What would you do if..."

How Decisions Get Made

After the onsite, each interviewer writes independent feedback. This typically includes:

A summary of the problem and the candidate's approach
Specific examples of positive and negative signal
A hire / no-hire / lean-hire / lean-no-hire recommendation

These writeups go to a hiring committee or a debrief meeting. The process varies:

Google:    Independent hiring committee reviews all feedback.
           Interviewers do not make the decision.
Amazon:    Bar raiser (experienced interviewer from another team)
           has veto power. Team makes the decision in debrief.
Meta:      Hiring committee reviews, then a separate team
           matching process happens.
Startups:  Usually the hiring manager + team lead decide in
           a meeting the same week.

The key insight: your interviewer is not the decision maker at most large companies. They are writing a report that someone else will read. This means your performance needs to be documentable. Vague positivity ("seemed smart") does not help. Specific signal does ("identified the O(n^2) bottleneck without prompting and optimized to O(n log n) using a balanced BST").

Timeline

The typical end-to-end timeline from application to offer:

Week 1:     Apply, resume screened
Week 2:     Recruiter call
Week 3:     Phone screen
Week 4-5:   Onsite scheduled
Week 5-6:   Onsite conducted
Week 6-7:   Hiring committee review
Week 7-8:   Offer extended (or rejection)

This can compress to 2 weeks at startups or stretch to 3 months at slower-moving companies. Having competing offers accelerates everything.

Company-Specific Variations

Not every company follows the template above. Some notable differences:

Take-home assignments: Some companies (especially smaller ones) replace the phone screen with a 2-4 hour take-home project. These test real-world coding more than algorithm knowledge. The upside is that you work in your own environment. The downside is the time investment — a 4-hour take-home for a company you are not sure about is a real cost.
Pair programming: Some companies (Pivotal, Shopify, others) do pair programming instead of whiteboard coding. You work on a real or realistic problem with the interviewer. This format rewards collaboration and practical engineering over algorithm recollection.
No algorithm rounds: A growing number of companies are dropping pure algorithm questions in favor of practical coding exercises (build a small feature, debug existing code, review a PR). These tend to be better predictors of on-the-job performance but are harder to standardize across interviewers.
Trial days or weeks: Some companies offer paid trial periods where you work on actual company code. This is the highest-fidelity evaluation method, but few companies invest in it because it is expensive to run.
AI-assisted coding rounds: Some companies now allow or even expect you to use AI coding tools during the interview, mirroring how work actually gets done. The focus shifts from syntax recall to problem decomposition, code review skills, and knowing what to ask the tool.

Negotiation Leverage Through Process

Understanding the process gives you negotiation power. Key tactics:

- Apply to multiple companies simultaneously so onsites overlap
- Share competing timelines with recruiters ("I have an offer
  deadline in 2 weeks") — they will expedite
- Ask recruiters for the compensation band before the onsite,
  not after the offer
- Request extra time for decisions; most "exploding offers"
  can be extended if you ask
- Remember that the recruiter wants you to accept; they have
  invested time and have a fill target

The companies with the most structured processes (Google, Meta, Amazon) are also the most predictable. Use that predictability to plan your preparation and timeline.

Common Pitfalls

Treating every stage the same: The recruiter call is not a technical interview. The phone screen is not the onsite. Adjust your preparation and energy accordingly.
Ignoring the behavioral round: Engineers consistently underprepare for behavioral interviews, then bomb them. At Amazon, a bad behavioral performance can sink an otherwise strong technical showing.
Not asking questions: Every stage ends with "do you have questions for me?" Having no questions signals low interest. Prepare 3-5 questions per interviewer.
Misunderstanding the hiring committee: At companies with committees, a single "strong hire" from one interviewer does not guarantee an offer. You need consistent positive signal across all rounds.
Applying cold: A referral from a current employee is the single most effective way to get past the resume screen. Use your network before applying online.

Key Takeaways

The interview process is a funnel designed to cheaply filter at each stage. Understand what each stage selects for and prepare accordingly.
Your interviewer is usually writing a report, not making the decision. Make your strengths easy to document with specific, concrete examples.
Behavioral interviews matter as much as technical ones. Underpreparing for them is the most common mistake experienced engineers make.
The process takes 4-8 weeks end-to-end. Plan your job search timeline around this, especially if you need to coordinate multiple offers.
Referrals bypass the most arbitrary filter (resume screen) and are worth pursuing before submitting any application.