Skip to content
Permalink
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
272 lines (224 sloc) 11 KB

Over the last 7 years I've interviewed about 100 candidates for various JavaScript roles, ranging from standard React/Angular/Vue frontend roles to database driver development to full stack logistics optimization. Most of the JavaScript interviewing guides out there overemphasize topics that are nice to learn but rarely matter in practice, like prototypical inheritance and functional programming. Here's the 3 questions that I usually use:

The Standard Junior Frontend Test

My first priority when considering a hire is whether they can code. If I'm hiring someone as a JavaScript developer, I want to know they can write JavaScript, not just talk about writing JavaScript. These interview questions are usually a follow-up to a code test.

There are two reasons why I do coding interviews even after doing a code test. First, you can find an answer for most questions online, and it's harder to Google your way through a short in-person interview. Second, I want to see the candidate's thought process and see how they work their way through a problem. Seeing the final product of their work can only tell you so much.

In my experience, as a frontend developer, you need to be able to load data from a REST-ish API and render it. Seems pretty basic, but you'd be surprised how many developers struggle with it. Here's the problem prompt I put on collabedit:

/**
 * Alpha Vantage exposes an API endpoint that lets you get the current price
 * of a stock. For example, if you make the below request, you get the current
 * price of Microsoft stock (MSFT).
 *
 * GET https://www.alphavantage.co/query?function=GLOBAL_QUOTE&symbol=MSFT&apikey=demo
 *
 * The response looks like `sampleResult` below.
 *
 * Implement a component `StockPriceComponent` that displays the current price
 * of a stock passed in to the component as `props.symbol`.
 *
 * You may use whatever HTTP client you prefer - fetch, Axios, etc. You should
 * not use Redux for the purposes of this exercise.
 */

import React from 'react';

class StockPriceComponent extends React.Component {

}
 
const sampleResult = {
  "Global Quote": {
    "01. symbol": "MSFT",
    "02. open": "139.3900",
    "03. high": "140.4200",
    "04. low": "138.6700",
    "05. price": "139.7850",
    "06. volume": "28338704",
    "07. latest trading day": "2019-10-24",
    "08. previous close": "137.2400",
    "09. change": "2.5450",
    "10. change percent": "1.8544%"
  }
};

TLDR; given a stock symbol like 'MSFT' or 'AAPL', write a React component that pings a REST API and displays the current price. React is a placeholder - I've asked the same question with Vue and Angular.

The API data is a little ugly, but that's part of the test. Tasks like accessing nested properties and properties with '.' in the name happen in practice.

Here's a basic solution in React using Axios.

const React = require('react');
const axios = require('axios');

class StockPriceComponent extends React.Component {
  constructor() {
    super();
    this.state = { status: 'LOADING' };
  }

  componentDidMount() {
    const symbol = this.props.symbol;
    const url = `https://www.alphavantage.co/query?function=GLOBAL_QUOTE&symbol=${symbol}&apikey=demo`;

    axios.get(url).
      then(res => this.setState({
        error: null,
        status: 'DONE',
        price: res.data['Global Quote']['05. price']
      })).
      catch(err => this.setState({
        error: err.message,
        status: 'ERROR',
        price: 'N/A'
      }));
  }

  render() {
    if (this.state.status === 'ERROR') {
      return (<h1>Error: {this.state.error}</h1>);
    } else if (this.state.status === 'DONE') {
      return (<h1>Price: {this.state.price}</h1>);
    } else {
      return (<h1>Loading...</h1>);
    }
  }
}

Depending on how well the candidate did, there's several follow up questions:

  • Error handling. How did the candidate surface errors? If using React, do they know about error boundaries?
  • Cleaning up resources. Suppose instead of displaying the price once, you want to ping the API every 5 seconds to update the price. How do you ensure the component cleans up after itself when it unmounts?
  • React hooks. How would this code look if written with hooks?
  • Promises. What happens if the response body changes? How would you rewrite this using Suspense?

The Async/Reactive Programming Test

Frontend framework tests are handy, but for full stack or backend devs I usually prefer to go into more detail about async, promises, and reactive vs imperative rather than getting lost in the minutiae of frontend frameworks. Here's the prompt:

/**
 * A Node.js readable stream (https://nodejs.org/api/stream.html#stream_readable_streams)
 * can be thought of as an event emitter that emits 3 events:
 *
 * - 'data' when a new chunk of data is available, `stream.on('data', chunk => {})`
 * - 'error' when an error occurred, `stream.on('error', err => {})`
 * - 'end' when the stream has successfully completed, `stream.on('end', () => {})`
 *
 * You may assume that the `chunk` passed to the 'data' event handler is a string.
 *
 * Write a function `streamToPromise()` that, given a stream, returns an ES6
 * promise that resolves to the concatenation of all chunks emitted in 'data'
 * if the stream successfully completes, or rejects with the error emitted
 * in 'error' if there was one.
 */

function streamToPromise(stream) {
  return Promise.reject(new Error('not implemented'));
}

TLDR; given a simplified stream (no backpressure, etc.), return a promise that fulfills with the concatenation of all 'data' events emitted by the stream.

One thing I love about this question is that I've had to write this code dozens of times. There's always some odd npm module like ftp that insists on only providing a stream-based API when I really only want a promise-based API. Naturally, this wouldn't work for large files, but it works fine for the small CSV files that I use ftp to download. You can hear more about that on my latest appearance on the JavaScript Jabber podcast.

Here's the solution:

function streamToPromise(stream) {
  return new Promise((resolve, reject) => {
    let res = '';
    stream.on('data', data => { res += data; });
    stream.on('error', err => reject(err));
    stream.on('end', () => resolve(res));
  });
}

There isn't much code to write, but a lot of thought and understanding goes into writing this. The only potentially tricky bit of syntax is remembering what the Promise constructor looks like. But the most common cause of confusion when working through this question is reactive programming: you can easily tell the interviewer's JavaScript is rusty when they start writing a while() loop to try to exhaust the stream imperatively.

Here's a few follow-up questions that illustrate some of the details:

  • Suppose you have a malformed stream that emits an 'error' event after an 'end' event. What happens to the promise?
  • Is it possible for the above function to miss a 'data' event?
  • Can you process a stream imperatively? Why is doing so a bad idea?

The Node + MongoDB Test

Dealing with data that's too big to fit in memory is a common task in Node. In my experience data usually fits into memory, hence the streamToPromise() exercise, but even if you're not working with a huge data set, you'll likely see code that's designed to handle huge data sets. This exercise is about iterating through a MongoDB cursor.

/**
 * A MongoDB cursor can be thought of as an object that exposes a single
 * function `next()`. The `next()` function returns a promise that resolves
 * to the next document in the query result, or `null` if there are no
 * more documents.
 * 
 * Suppose you have a collection of stock holding documents in MongoDB
 * that look like this:
 * 
 * { symbol: 'AAPL', shares: 10, userId: 123 }
 * 
 * Given a `cursor` over this collection, write a function that calculates
 * the average value of a user's portfolio.
 * 
 * You may assume the `getPrice()` function asynchronously calculates the
 * current price of a given stock symbol.
 */

const getPrice = require('./getPrice');

async function calculateAveragePortfolio(cursor) {
}

TLDR; given a collection of stock holdings that do not include the current price, calculate the average value of a user's holdings.

At a high level, this looks a lot like the React StockPriceComponent question, but geared towards backend. Depending on the candidate's experience, I may ask them to implement getPrice() as well. Here's a few concepts this exercise tests:

  • Collecting unique values
  • Basic caching of requests
  • Comfort level with promises and async/await. Can the candidate actually iterate a cursor with async/await?

Below is the solution:

const getPrice = require('./getPrice');

async function calculateAveragePortfolio(cursor) {
  const prices = {};
  const userIds = new Set();
  let totalValue = 0;

  for (let doc = await cursor.next(); doc != null; doc = await cursor.next()) {
    let price = 0;
    if (prices.hasOwnProperty(doc.symbol)) {
      price = prices[doc.symbol];
    } else {
      price = await getPrice(doc.symbol);
      prices[doc.symbol] = price;
    }

    userIds.add(doc.userId);
    totalValue += doc.shares * price;
  }

  // Optimization: no need to calculate the size of each individual user's
  // portfolio if all you need is the average. You just need the sum of all
  // the portfolio sizes and the number of users.
  return totalValue / userIds.size;
}

Here's some follow-up questions:

Moving On

Interviewing is a touchy subject. These are the questions that I've used to screen JavaScript developers for the last several years. Take them with a grain of salt - I guarantee you that other people use different interview questions. However, I think these interview questions best capture the skills that I've found most valuable in my own day-to-day work. Feel free to try them out, whether you're an interviewer looking for inspiration for your own questions or an interviewee looking for practice problems.

You can’t perform that action at this time.