Practice a real company interview
Pick the job you're interviewing for and step into a live voice mock calibrated to what it actually tests.
Software Engineer
Amazon's new-grad SWE loop is roughly one online assessment plus four to five interviews, and it is famous for weighting behavioral as heavily as coding. Every round pairs a LeetCode-style problem with deep Leadership Principles questioning, so you are graded as much on how you make decisions and own outcomes as on whether your code runs. Expect a bar-raiser round that probes for depth on one specific past project.
Practice this interviewData Scientist
Databricks' new-grad Data Scientist loop spans four core areas: SQL/data manipulation, applied statistics and A/B testing, machine-learning fundamentals, and a coding round, usually wrapped with a behavioral and a case-style discussion. It is genuinely broad — you're expected to be fluent in querying data, reasoning about experiments, and explaining models from first principles, not just calling a library. Expect to defend your statistical choices out loud.
Practice this interviewSoftware Engineer Intern
Datadog's SWE intern loop typically starts with a HackerRank OA, then one or two technical interviews plus a behavioral conversation. The coding lean is practical over trick-question — clean data-structure problems and, often, a small hands-on task — reflecting a company built on high-scale observability. Interviewers probe genuine curiosity about systems, monitoring, and data, and they care whether you can reason about what happens at scale.
Practice this interviewSoftware Engineer
Google's new-grad loop is typically two phone/virtual screens followed by an onsite of four-plus interviews, and it leans harder on raw algorithmic depth than almost any other company. Problems trend toward medium-hard, and interviewers care as much about how you reason toward an optimal solution and analyze complexity as about the final code. "Googleyness & Leadership" is scored separately, and hiring is decided by a committee reviewing your written packet, not the interviewers in the room.
Practice this interviewSoftware Engineer
Meta's university-grad loop is built around speed and breadth: two "Ninja" coding rounds where you're expected to solve two medium problems in ~35 minutes with correct, running code, plus a behavioral round and (for some) a lightweight design discussion. Interviewers value clean solutions delivered fast over drawn-out perfection, and there's a strong emphasis on communicating trade-offs. Team matching happens after you pass, so the loop tests general engineering strength, not one stack.
Practice this interviewSoftware Engineer
Microsoft's new-grad loop is usually an OA followed by three to four interviews, often ending with an "as-appropriate" round with a senior engineer or manager. The coding bar is fair — think solid easy-to-medium problems — but interviewers care a lot about how you think, whether you ask clarifying questions, and how you communicate. It's a more conversational, collaborative loop than the pure-algorithm gauntlets, with real weight on culture-fit and growth mindset.
Practice this interviewFull-Stack Engineer
Rill is an early-stage startup, and its full-stack new-grad loop reflects that: a take-home or live build where you ship a small end-to-end feature, a round walking through your code and the trade-offs you made, and a scrappy behavioral chat about how fast you learn and ship. There's no algorithm gauntlet — they want proof you can move across the stack (frontend, API, data) and get something real working with limited hand-holding.
Practice this interviewBackend Software Engineer
Stripe's backend loop is known for being practical rather than puzzle-heavy: instead of abstract LeetCode, you get a hands-on coding round in a real editor (often building or extending a small service), an API/integration round, and a debugging round on an unfamiliar codebase. Because Stripe moves money, correctness, idempotency, and reliability are treated as first-class skills. Expect to reason about failure modes and edge cases far more than about clever asymptotics.
Practice this interview