Two rounds, one session

The rounds LeetCode skips

A failing test to debug, and a real-world case to design — with an AI interviewer that reacts as you work.

Watch a session, 60 seconds

codesparring.dev/interviewScripted demo

Interviewer

Array of prices, one target. Return the indices of the two entries that sum to it. Before you code, what do you want to ask?

# solution.py

3 tests ready · Run Tests

A scripted demo of the real session format: the interviewer reads your code, your tests actually run, and the round ends with a score.

What a session looks like

Bug Fixcheckout total returns NaN

✕ 1 failingcart.test.ts › applies discount

expect(total).toBe(90)  // Received: NaN
function applyDiscount(price, pct) {
  // pct arrives as "20", not 0.2
  return price - price * pct;
}

Before you change the formula, what are the actual types of price and pct?

Case LabPalantir · FDSE

Good scope. So what's the contract for the dispatch API, and which tradeoff are you defending?