Tech Interview Rounds Explained 2026
Every round in a tech interview loop decoded: what each round tests, how it is scored, and how to prepare for it specifically.
The Recruiter Screen
The recruiter screen is 20-30 minutes and non-technical. The recruiter is checking: are you actually interested in this role, do you meet minimum requirements, and is your compensation expectation in range? It is a filter round, not a scoring round. No one passes the recruiter screen and is then offered a job. It just moves you to the next stage.
What to prepare: a 90-second summary of your background, a clear answer on why this company (specific, not generic), and your compensation expectation. Research the role's level range before the call. Saying "I'm looking for L5 equivalent" is more credible than "open to whatever you think is fair."
Online Assessment (OA)
Online assessments are timed, asynchronous coding challenges. Standard format: 2 problems in 90 minutes, medium-difficulty LeetCode. Common for new grad pipelines at Amazon, Google, Microsoft, and most mid-size tech companies. OA results are often automatically processed by a scoring system before any human reviews them.
OA-specific strategies: read both problems before starting. If one is harder, do the easier one first to bank points. Write brute-force solutions that pass partial test cases rather than spending all your time on an optimal solution that doesn't run. Partial credit exists. An O(n^2) solution that passes 15/20 test cases scores higher than an O(n log n) solution with a bug.
Technical Phone Screen
The technical phone screen is a live 45-60 minute coding session with a senior engineer. Typically 1-2 medium problems. You share your screen or use a shared code editor. Unlike the OA, this round has a human interviewer and is scored on communication as heavily as correctness.
The phone screen format favors candidates who narrate their thinking. Before typing a single character, spend 2-3 minutes clarifying requirements, stating your approach, and confirming with the interviewer. Interviewers who are writing their feedback packet need evidence of your reasoning, not just your code.
System Design Round
System design rounds are 45-60 minutes, required at mid-level and above (L4/SDE-II). The interviewer gives you an open-ended problem: "Design Twitter's timeline", "Design a ride-sharing service." You whiteboard or diagram a distributed system architecture and discuss tradeoffs.
Scoring is based on: requirements gathering (did you clarify scope?), component thinking (can you decompose into services?), scale awareness (did you design for real load?), and trade-off articulation (can you explain why you chose Kafka over RabbitMQ?). Read the full System Design Round Guide for the complete framework.
Behavioral Round
Behavioral rounds use STAR-format questions about your past work: "Tell me about a time you dealt with conflict on a team", "Describe a project that failed and what you learned." Every company has at least one behavioral round. Amazon has five or six.
The mechanical fix for behavioral rounds: prepare 6-8 specific STAR stories covering leadership, conflict, failure, ambiguity, and scope expansion. Write them down before your interview. The stories you have prepared will be 80% of what you actually say, so having them ready matters more than improvising well under pressure.
Bar Raiser and Final Design Rounds
Amazon's Bar Raiser is the most well-known final-calibration round, but Google's Hiring Committee review and Microsoft's "As Appropriate" interview serve similar functions. These rounds exist to calibrate your seniority against the existing team, not to test new skills. The interviewer is asking: "Would this person raise the bar compared to the median engineer we'd hire?"
Preparation for calibration rounds: be specific about scope. Vague claims ("I led a large-scale initiative") score worse than specific ones ("I designed the migration plan for moving our 50-service monolith to Kubernetes, reducing deploy time from 4 hours to 12 minutes"). Numbers and specificity signal seniority.
Online Assessment (OA) Round
Many companies (notably Amazon, Microsoft, and most internships) start with an Online Assessment: 60-90 minutes, 2-4 algorithmic problems on a platform like HackerRank or CodeSignal. These are auto-graded with test cases.
OA strategy: read all problems first, attempt the easiest one to lock in points, then tackle harder ones. Many OAs use partial credit per test case, so a brute-force submission that passes 60% of tests can advance you. Reports on LeakCode show Amazon OAs in 2025-2026 trend toward two medium problems (1 array, 1 graph or DP) with about a 70% advance threshold.
Hiring Manager Round
Most loops include a hiring manager round, typically at the end. It blends behavioral with team-fit questions. The hiring manager will ask about your career goals, why this team specifically, and how you handle their team's specific operational realities (on-call cadence, deployment frequency, code review culture).
This round has hidden veto power at most companies even when the formal decision is committee-driven. A strong hiring manager fit can carry an ambiguous loop; a weak hiring manager fit can sink a strong loop. Prepare specific questions about the team and demonstrate genuine interest in their work. Reports on LeakCode tag "HM round" as the single most predictive round of final outcome.
See What Each Round Asks
Browse real interview questions by round type from verified candidate reports.
Browse by Company and Round