On tech interviews
After doing a lot of interviews during two job changes in the last two years I picked up a thing or two about interviewing, collectively between sending résumés, having calls, coding exercises and technical interviews, as well as interviewing people myself, I racked up over 100 'interview stages', and I have been approached by many people about tips on interviewing, this post summarizes most of the things I talk about.
There are plenty of articles out there bashing technical interviews for being excessively complicated at often including unrelated exercises to the skills the job position you are applying for. Make no mistake, interviewing is it's own skill, and since changing jobs is one of the fastest way to increase your earnings, it can be a skill worth millions of dollars as evidenced by many businesses addressing the need to interview prep.
Usually the first step after getting your CV noticed is to schedule a call with a recruiter, it lasts roughly 30 minutes and it's unlikely that there will be technical questions asked unless stated previously.
During the initial call the recruiter will probably give further information about the position and the company, they might also ask you for an introduction and discuss motivation, knowledge and expectations
This might be a good time to get your first questions in, screen about salary expectations, how long would it take to get feedback on each stage, and other general company information you are concerned about, if it's a startup, information on growth, financial health and recent developments can be very important.
As a rule of thumb you should always have questions prepared at every step of the interview, it allows you to gather more information and it shows interest. It is also useful to read the job posting again a few hours before each interview to avoid unnecessary questions.
Many companies will present some coding challenge to a candidate to test coding ability before taking more time to interview them, usually with a platform like Codility or HackerRank. Most of the times these tests will be some simple logic/algorithm challenge and won't involve much complexity, usually with a long form text with examples for input and expected output that serve as test cases, often they can be automatically graded and won't be reviewed by a human interviewer.
These tests will usually have some restrictions on what programming language or environment you can code in, sometimes you just need to complete some snippet while others you might have to parse input from STDIN, they are often timed, from 1-3 hours depending on complexity, most HackerRank tests include three problems of varying difficulty and some companies might require you to solve only 2 out of 3.
#Take home assignments
Similar to coding challenges, take home assignments are handed as a way to scan candidates before dedicating more interviewer time, however they are usually expected to be done in days or even a week, they usually consist of a practical task which includes building a working application including tests, some companies will still request you to timebox up to a few hours and see how far you can go with it.
These are great to showcase different types of knowledge in application development, from writing tests, to focus on user interface and usability, coding standards and organization, there are many ways reviewers can evaluate your code, often there will be restrictions on what languages or frameworks you can use, and once again, in case there isn't prefer to use something the company utilises or that is widely known instead of some obscure tech, I've seen many candidates trying to use different tools to try and trick reviewers who are not as likely to understand the code.
One important thing to note here is that it's not just code that counts, reviewers might not notice but you can stand out by writing good and detailed commit messages, a detailed README with clear instructions how to run the application and its features, and informative comments that help navigate the code, even deploying the application somewhere to be quickly tested is a plus.
Whiteboarding is one of the most feared types of interview by developers, it involves writing code in a whiteboard with a pen instead of using an editor or an IDE, problems are usually of algorithmic nature and are dictated by the interviewer at the beginning of the interview and timed between 30 minutes to an hour.
Lots of people get very nervous when doing these types of interviews, the best way to be prepared is to just practice on pen and paper yourself, some people will even practice on a real whiteboard but I think that's not necessary, it might make you better at using the space available, however.
A big mistake many developers will commit is to start writing on the board right away to save time, what you should do instead is repeat the problem out loud in your own words to communicate with your interviewer and to ask questions before jumping in, most of these interviews expect you to dig around the problem to understand edge cases, limitations or solution sets.
It is also important to choose a widely known programming language, Python is a good choice because it requires you to write less, it helps picking a language most developers know because interviewers can jump in and ask details about constructs as well as suggest potential fixes on the go.
The best whiteboard interviews I've had are the ones that start with a simple premise so you can get some code on the board and increase in complexity as the interviewers add further exercises to the problem, expecting you to adjust your code as needed. This way you'll be judged on how far you've gotten and you don't necessarily have to have solved everything.
The biggest companies like Google, Amazon and Facebook are famous for these types of interviews, many of them already allow you to type your solutions in a simple editor instead of a whiteboard, in any case, if you still face a whiteboard, some people recommend you to bring your own pens to not have to waste time having to finding working ones at the interview site.
I feel systems design is one of the most difficult types of interview out there, as they are usually quite open ended and varying in complexity. They are often overlooked as most people focus on getting their algorithms right, I only found a few YouTube videos and a handful of articles as study material for systems design.
In systems design interviews you'll be asked to draw on a whiteboard the components of a hypothetical system that is often inspired by a real life service like Twitter or OpenTable, these components can range from interface design, database design, interactions between services and third parties, asynchronous operations and consistency, and many other traits.
Just like an algorithms whiteboard interview, it's important to keep it as a conversation, often bouncing ideas and questions from the interviewers to discuss possible solutions and understand what they are looking for, there is usually not a single definitive answer.
Most systems design questions in web applications will often involve understanding scalability for up to millions of users, so brush up on your understanding of horizontal and vertical scalability, load balancers, async operations, database design and replication.
You might also get asked about your own experiences with these different technologies, you might as well come prepared with anecdotes described in the STAR model, which is a good way to be able to showcase what you've done so far that might be relevant to your interviewers.
#General tips & Resources
Always get feedback from your interviewers, especially if you didn't get an offer, so you can improve in the future. If you are applying for multiple roles it might be useful to keep a spreadsheet with all your applications where you can gather all feedback
Asking your interviewers for a time frame when you might hear back is a good way to reduce anxiety, so you can ping them if they don't respond within the agreed time frame.
Always ask questions at every single stage of the interview, do your homework on the company to be able ask more relevant questions.
Some people like to do practice or mock interviews and there are service which you can use for that, however, the best practice is to get real interviews with real stakes and get some experience.
Cracking the Coding Interview is probably the most well known book out there to practice algorithm questions, it is one of the best too, it is succinct and complete with all you need to know about algorithms and data structures for interview, you should also do some practice problems in one of the many coding challenge websites out there, my favorites are HackerRank and LeetCode.
I will try to keep this post updated the more I become knowledgeable about the topic, for now, I hope you enjoyed it, and good luck in your interviews! If you found this post useful I would love to hear feedback about it, you can reach me on twitter @mgrisoli.