AI Technical Interview Coach
Mockloop runs a full FAANG-style technical interview, tracks the exact concepts you trip on, and uses spaced repetition to re-test you on your real blind spots — not a random problem queue.
Traditional platforms like LeetCode and HackerRank evaluate code correctness, not pedagogical understanding. They can tell you when a test case fails, but they don't act as a tutor. They don't remember the specific conceptual gaps you have — recursion base cases, off-by-one errors in sliding windows, DP state transitions — and they don't explicitly drill you on those blind spots.
Mockloop is built around the opposite premise: an interviewer that learns you over time and actively coaches you through the parts you don't yet understand.
Three things set Mockloop apart from auto-graders and problem banks.
The AI maintains a living memory file of your progress. It tracks the exact conceptual gaps, syntax errors, and algorithmic misunderstandings you make across different coding sessions — so next week's interview picks up where last week's left off.
Instead of randomly selecting problems or blindly grinding, Mockloop uses spaced-repetition intervals to explicitly re-test you on concepts you previously struggled with. The things you almost understood don't slip through the cracks.
Rather than just grading your code, the AI acts as a mentor. It forces you to explain your approach, hands you counter-examples when your algorithm is flawed, and coaches you to fix the root cause of your bugs — curing patch-driven debugging.
Split-screen workspace mirroring real interview environments: a Monaco code editor on one side, a chat-and-whiteboard channel on the other. Here's how a session runs.
You ask for a problem. The AI drops an unassisted problem description into the editor — and explicitly refuses to provide starter code, function signatures, or hints. You're on your own, exactly like a real interview.
You ask clarifying questions in the chat. The AI adopts an Interviewer persona to answer realistically, and a Coach persona to privately evaluate whether you asked about the right edge cases. Missed a critical constraint? That goes in your memory file.
You explain your algorithmic approach. If there's a flaw, the AI won't correct you directly — it will hand you a counter-example so you can discover the flaw yourself. This is how staff-level interviewers actually work.
Once you write code, the AI silently reads your editor state, reviews your idiomatic fluency, demands time and space complexity justification, and suggests optimizations — the full rigour of a staff-engineer review.
After the session, your specific algorithmic gaps ("struggles with DP state variables", "misses off-by-one in sliding windows") are logged into the knowledge graph. Your spaced-repetition queue is updated for next time.
Sign in with Google and run your first mock interview in under a minute.
Start a mock interviewMockloop is an independent project built by an engineer who got tired of grinding problems without improving on the concepts that actually mattered. It's designed for software engineers preparing for technical interviews at top-tier companies — but the pedagogical approach (memory + spaced repetition + active coaching) works equally well as a general-purpose algorithms tutor.