Technical assessment platform

Hire engineerswho can build.

Real IDE. Real problems. AI scoring that shows you how candidates think — not whether they memorized the right algorithm.

Start for freeSee how it works
litmus.app/interview/sess_8f2k9…
auth.ts — line 23
export async function verifySession(token: string) {
const parts = token.split(".");
const payload = JSON.parse(atob(parts[1]));
return payload.exp > Date.now() / 1000;
}
⚠ Flaw detected: JWT signature is never verified — tokens can be forged.

Process

Five competencies. Two components that measure them — and one popular approach that doesn't.

Bug Bash

See how candidates debug when their AI tool gets it wrong.

What we learn

  • ◆Critique
  • ◆Skepticism of AI output

System Design

See how candidates navigate ambiguity and adapt to evolving requirements.

What we learn

  • ◆Decomposition
  • ◆System reasoning at scale
  • ◆Product thinking & ownership
Misses everything

LeetCode / HackerRank

Score 0 across the board. Pattern matching isn't engineering.

Why Litmus

Built different, on purpose.

Most technical screens measure interview performance. Litmus measures engineering judgment.

A real development environment

Candidates work in a full VS Code-style editor with file tree, terminal, port forwarding, and AI tooling — the same stack they'd use on the job. Not a textarea. Not a CodePen.

Designed to defeat AI shortcuts

Our flaw-finding assessments seed the codebase with bugs the AI will confidently misdiagnose. Candidates who just ask the AI to fix the failing tests will apply hallucinated patches that pass the visible tests but quietly fail the hidden ones — so you see who's actually reading the code.

Scoring that explains itself

Every submission is graded against the five engineering competencies that matter — critique, AI skepticism, decomposition, scale reasoning, and product ownership. Each result ships with per-flaw evidence and per-criterion reasoning, not just a number.

Send an assessment in minutes

Create a plan, generate an invite link, and share it. Candidates start within their browser — no accounts, no installs, no scheduling. Results land in your dashboard within ~1 hour of submission.

"We replaced our take-home with Litmus and cut time-to-offer in half. The flaw scoring actually told us something — for the first time, we could explain why we passed on a candidate."
JR

Jamie R.

Head of Engineering · Series B startup

Stop guessing.
Start knowing.

Create an account and send your first assessment in minutes. No contracts, no setup calls.

Get started free