What each interview loop tests, role by role: frontend, data science, data engineering, machine learning, DevOps and SRE, security, QA and SDET, product management, marketing and new grads. For each you get what gets tested, four questions to expect, what a strong answer shows and where to focus your preparation, with links to a longer question list for the role. For language questions, see interview help by programming language.
How to use this page
Every role below has a section with what gets tested, four questions to expect, what a strong answer shows and where to focus. Most loops also include a coding or analysis round and a behavioral round, so pair this page with our coding interview help and behavioral interview help.
Engineering roles
Frontend engineer
What gets tested: JavaScript fundamentals (closures, the event loop, promises and this), a framework such as React (state, effects, rendering and keys), CSS layout, browser performance and Core Web Vitals, accessibility, and frontend system design.
Questions to expect:
- Implement debounce, or Promise.all, from scratch.
- Why does this component re-render, and how would you stop it?
- Design an autocomplete component used by millions of people.
- How would you make this page accessible to a screen reader user?
What a strong answer shows: working code in the browser plus the reasons behind it: performance, accessibility and user experience.
Prep focus: practice two UI coding tasks and one component design; see frontend system design and JavaScript interview questions.
DevOps and SRE
What gets tested: Linux, networking (DNS, TCP, HTTP and load balancing), containers and Kubernetes, CI/CD, infrastructure as code, a major cloud platform, observability and SLOs, and incident response, often through troubleshooting scenarios.
Questions to expect:
- A service is suddenly slow. Walk me through what you check, in order.
- What happens when you type a URL into a browser?
- How would you roll out a risky change safely?
- Explain the difference between a liveness probe and a readiness probe.
What a strong answer shows: a calm, systematic method: form a hypothesis, check the cheapest signal first, and say what you would change to stop it happening again.
Prep focus: practice two troubleshooting stories from your own work; see DevOps interview questions and Kubernetes interview questions.
Security engineer
What gets tested: application security (injection, cross-site scripting, request forgery, server-side request forgery and broken access control), threat modeling, cryptography basics, cloud security and least-privilege access, incident response, and secure code review.
Questions to expect:
- How would you threat-model a password reset flow?
- Find the vulnerability in this code snippet.
- What is the difference between hashing and encryption, and when do you use each?
- A key was leaked in a public repository. What do you do in the first hour?
What a strong answer shows: that you think like an attacker and fix like an engineer: likely attack paths first, then practical fixes.
Prep focus: prepare one incident story and one threat model; see security engineer interview questions.
QA engineer and SDET
What gets tested: test design (equivalence classes and boundary values), the test pyramid, automation frameworks, API testing, CI integration and flaky tests, bug reporting, and coding for SDET roles.
Questions to expect:
- How would you test a login page?
- Which tests would you automate first, and which would you never automate?
- A test fails one run in ten. How do you deal with it?
- Write a test for this function, including the edge cases.
What a strong answer shows: risk-based thinking: what matters most to users, what is most likely to break, and what is cheapest to check.
Prep focus: practice one 'how would you test it' question per day; see QA engineer interview questions.
Data and machine learning roles
Data scientist
What gets tested: statistics and probability (hypothesis tests, p-values, power and A/B testing), SQL, machine learning basics (bias and variance, regularization, metrics), product sense, and analysis case studies, usually in Python.
Questions to expect:
- Our A/B test shows a 2% lift with a p-value of 0.06. What do you recommend?
- How would you measure the success of a new feature?
- Explain precision and recall to a product manager.
- Write a query for seven-day retention by signup week.
What a strong answer shows: statistical reasoning tied to a business decision, not just the formula.
Prep focus: rehearse one experiment design and one metrics case aloud; see data scientist interview questions.
Data engineer
What gets tested: advanced SQL, data modeling (star schemas and slowly changing dimensions), batch and streaming pipelines, orchestration, data quality, warehouses and partitioning, and Python for transformations.
Questions to expect:
- Model the tables for an online store's orders and returns.
- How would you rebuild a day of data after a bug, without duplicates?
- Batch or streaming for this use case, and why?
- How do you know a pipeline produced correct data?
What a strong answer shows: pipelines that are idempotent, observable and cheap to rerun, with data quality checks built in.
Prep focus: draw one pipeline end to end with its failure modes; see data engineer interview questions.
Machine learning engineer
What gets tested: machine learning fundamentals and evaluation, feature engineering, deep learning basics, ML system design (ranking, recommendations, fraud), MLOps (training and serving skew, monitoring and drift), and increasingly LLM topics such as retrieval, fine-tuning and evaluation.
Questions to expect:
- Design a system to rank items in a home feed.
- Your model's offline metrics improved but online engagement fell. Why?
- How would you detect data drift in production?
- When would you fine-tune a model rather than use retrieval?
What a strong answer shows: that you can take a model from data to production and keep it working, not just train it.
Prep focus: prepare one ML system design and one project deep dive; see ML system design and ML engineer interview questions.
Product and marketing roles
Product manager
What gets tested: product sense (design a product for a group of users), execution and metrics (what would you measure, and why did a metric drop), prioritization, strategy, and behavioral questions about leadership without authority.
Questions to expect:
- Design a product to help people find a roommate.
- Daily active users fell 10% last week. What happened?
- How would you prioritize these five requests?
- Tell me about a time you disagreed with engineering.
What a strong answer shows: a clear user, a clear problem, a reasoned choice between options, and how you would know it worked.
Prep focus: run one product-sense and one metrics case aloud each day; see product manager interview questions.
Marketing
What gets tested: channel and growth strategy, campaign case questions, metrics such as customer acquisition cost, lifetime value, return on ad spend and conversion rate, positioning and messaging, experiments, and walkthroughs of past campaigns; take-home assignments are common.
Questions to expect:
- How would you launch this product with a small budget?
- A campaign's cost per acquisition doubled. What do you check?
- Walk me through a campaign you ran and what you would change.
- How would you position us against a larger competitor?
What a strong answer shows: numbers tied to decisions: what you spent, what it returned and what you learned.
Prep focus: prepare two campaign stories with real metrics; see marketing manager interview questions.
New grads
New grads
What gets tested: data structures and algorithms, your projects and internships in depth, behavioral questions about teamwork and learning, and online assessments early in the process, usually on a fixed recruiting timeline.
Questions to expect:
- Walk me through the hardest bug in your project.
- Tell me about a time you worked with a difficult teammate.
- Solve this array problem, then improve it.
- Why this team, and what do you want to learn?
What a strong answer shows: clear explanations of your own work and steady problem solving, not years of experience.
Prep focus: practice online assessments under time limits and two project walkthroughs; see new grad online assessments and intern interview questions.
Language-specific questions for these roles are on interview help by programming language.
Where CoPilot Interview fits
The quickest way to find the gaps in your preparation is to answer real questions out loud and hear how they land. CoPilot Interview is an AI interview assistant; its free AI mock interview builds role-specific questions from your résumé and the job description, reads them aloud, scores each spoken answer out of ten and shows a model answer to compare. The free plan gives 15 minutes a day of live help in sessions of up to 15 minutes, with AI mock interviews included, and the browser demo at /try has 145 practice questions with no signup. If you are considering live help in a real interview, follow your employer's rules, and remember that online assessments and AI-led screens usually forbid it.
Disclosure: CoPilot Interview is our product, so weigh what we say about it accordingly.
Frequently asked questions
Do all tech roles have a coding round?
Most engineering, data and machine learning roles do. Product and marketing roles usually replace it with cases, metrics questions or a take-home assignment.
How many behavioral questions should I prepare?
Six to eight stories cover most loops: a success, a failure, a conflict, a time you led without authority, a time you changed your mind, and a technical or analytical challenge. Each story can answer several questions.
How long should I prepare for a role-specific loop?
Two to four weeks of focused practice is typical if you already work in the field, longer if you are switching roles. Spend most of it answering out loud rather than reading.
Are take-home assignments common?
In data, marketing and some frontend roles, yes. Treat the write-up as part of the test: state assumptions, show your reasoning and keep it short.
Where are the old role help pages?
Each role that used to have its own help page now has a section here, and the old addresses redirect to the matching section.
Practice this interview before the real one
The AI mock interview builds questions from your resume and the job description, reads them aloud and scores each spoken answer out of 10. Included on the free plan.
Start a mock interview