Amazon Software Developer Interview Questions & Answers
Below is a list of our Amazon interview questions. Click on any interview question to view our answer advice and answer examples. You may view 15 answer examples before our paywall loads. Afterwards, you'll be asked to upgrade to view the rest of our answers.
1. How would you define software configuration management?
How to Answer
In simple terms, software configuration management is the process of systematically managing, tracking, and controlling the changes that happen in the software development lifecycle. This includes changes in the code, documentation, and process. Change control, however unpopular, is a function that ensures that all changes made to the software system are consistent with Amazon development practices and internal methodologies. The ultimate goal is productivity improvement with minimal mistakes throughout the development lifecycle. Here, the interviewer is looking for how you prioritize configuration management duties and responsibilities and your experience with controlling something even if you don't know what it is. Configuration management is used when multiple team members are working on software or a revision that is running on multiple machines or needs constant updating. It may also be necessary if a company has multiple locations or divisions that need to distribute work concurrently.
Answer Example
"My definition of software configuration management is a set of activities that identify several items including source code modules, test cases, requirements specification documents, tools used, objects, files, and CSCI in SCM repositories, just to name a few. There are also other important activities related to these actions. The first thing I look at is a baseline which lets me know the accepted version of a software configuration item. The second is change control, which is a method that ensures quality and consistency when changes are made. Third on this list is configuration audits which verify that the software meets the minimum baseline needs and that the deliverable is being met."
2. How much do you know about object-oriented programming?
How to Answer
There are two different types of programming. These include Object-Oriented Programming and Functional Programming. Knowing the differences between these two approaches and their benefits is critical for an experienced Amazon software developer. Interview questions are meant to probe your knowledge, experience, and qualifications for this position. Interviewers use a wide range of technical questions to help them accomplish this. This is an example of this. Technical questions are easily answered by defining the term you are being asked about and then discussing its features and benefits. You can also provide an example of how it is used to illustrate your answer.
Answer Example
"I am very familiar with Object-Oriented Programming, also known as OOP, and use it a lot. This type of programming defines the data structure and the types of operations applied to the data. I like this approach because it helps me organize my code logically around a specific parameter. It also makes it easier to maintain and document the code. Finally, objects can be inherited and extended, which accelerates and shortens the software development life cycle. Based on my research, I believe the software developers here at Amazon also prefer OOP."
3. What are some of the testing and debugging processes you use in your software development?
How to Answer
This is either a standalone question or a follow-up to a previous question. If you previously told the Amazon interviewer about the QA processes you use, you should now discuss your testing and debugging processes. During an interview, you should always anticipate follow-up questions whenever you answer an interviewer's question. This is why you should keep your answers short and direct. It creates an opportunity for the interviewer to dive deeper into the topic if they have a specific interest in it or move on to a new line of questioning.
Answer Example
"I typically use unit testing tools to identify errors in my code when I develop software. I follow this with a debugging process and additional testing of the bug fixes. This is known as White Box Testing. The debugging tools I use most often are GDB and DBX, which are both public domain products. I prefer these because they provide command-line interfaces. I used command-line when in school, and it has always been easier for me to work with than a GUI. Some other tools I use are code-based tracers, profilers, and interpreters."
4. What are some of the QA processes you are familiar with and have worked with in the past?
How to Answer
While a software developer may not be directly involved with the testing and quality assurance of the code they develop, they should be familiar with the processes. This will help them code for the requirements of Amazon. The interviewer uses this operational question to determine the type of quality control systems you have worked with and confirm that your processes are similar to the ones they use. Even if your experience isn't an exact match to Amazon's, you should be able to describe how you would quickly transition to their processes.
Answer Example
"Because I develop software using the Agile framework, the quality control processes I encounter the most are the ones used within this framework. I have worked with Waterfall, Just In Time, and SQA as well. I feel comfortable transitioning between QA processes, and I am confident I can easily adapt to the ones that the Amazon team is using."
5. Here at Amazon, we use Agile and Waterfall. Are you familiar with these software development methodologies and, if so, which of these do you prefer?
How to Answer
An interviewer at Amazon will ask this question to determine if you are a good fit for their organization. There are a lot of different methodologies software developers can choose from. Regardless of which one you prefer, you'll need to adapt to the software development methodology used by Amazon. You can learn about the methodology the company's software development team uses should through your pre-interview research. The methodology you state a preference for should align with the one Amazon currently uses.
Answer Example
"The software development methodology I prefer and use the most is Agile. The reason I like this iterative framework is its daily stand-ups. These provide the team with the opportunity to discuss its development goals and any issues that they've encountered. I prefer to use the Agile methodology because it identifies errors and enables them to be corrected at each stage of the process. The result of this is better quality software and a reduced need for debugging toward the end of the project."
6. Can you describe a challenging software development project you recently completed and the results?
How to Answer
Ideally, you have encountered some of the same challenges that Amazon has during your software developer experience. The interviewer hopes to confirm this. It demonstrates that you will address the challenges and problems they encounter and help them resolve these challenges quickly. Organizations like Amazon hire people to either fill vacant roles or expand their team. In either case, they expect the candidate to bring a fresh perspective along with skills, experience, and knowledge they currently don't have on their team. Your answer should affirm this.
Answer Example
"The most challenging software project I've worked on recently involved an application to allow the company's personnel to self-manage their HR benefits. The challenge was that end-users had to have secure access to their information without the HR staff intervening. This required me to implement several security features, including user authentication, password protection, and the ability for users to access the company's database behind its firewall via VPNs. Recognizing that I did not have recent experience in these technologies, I refreshed my skills by taking a short online course and collaborated with some of the other developers on our team on the project. Together we developed the software on schedule and under budget. I learned to recognize my shortcomings and bring in additional resources when needed."
7. What are some of the software design patterns that you use, and in what situations do you use them?
How to Answer
Amazon interviewers will ask you operational questions during an interview for a software developer's position to determine how you complete the tasks required by this role. They also want to learn that your techniques are similar to those of their current development team. It is best to respond to Operational questions succinctly and directly. You should always expect follow-up questions when an Amazon interviewer is asking about a specific task. This indicates that they have a special interest in the topic they are questioning you about.
Answer Example
"Throughout my career, I've tried several different software design patterns. Some of these were Singleton, MVC, and Iterator. However, the one I use the most is Template. This design pattern helps me code straightforwardly and accelerates the software development cycle. It also results in fewer errors and less debugging."
8. What are some of the types of development projects you have worked on in your past positions?
How to Answer
Amazon prefers to hire candidates who have direct experience with the type of work they do. An interviewer will ask this question to determine if you've recently done the type of work they want to hire you for. During an interview with Amazon, all your answers should address the needs of the employer. It would be best to discover these during your pre-interview research about the company, its products, and its practices. You can find the information you need in the job posting and on Amazon's website.
Answer Example
"In my last two roles, I was focused on improving applications that manage transactions and interact with relational database systems. My role in the application development was to develop all the code to control data queries, lookups, and indexing of the data. I worked with other programmers to design the interface, manage the backend database, and develop other application features. The companies I worked for were focused on e-commerce, conducting sales and other transactions via their websites, much like Amazon."
9. I've heard applications referred to as both software and programs. Which terminology do you use, and how do these differ?
How to Answer
During a software developer interview at Amazon, you will be asked to discuss terms, processes, and other aspects of your job. An interviewer will ask these questions to explore your knowledge, skills, experience, and qualifications for this job. When presented with two similar but different terms, you should be prepared to compare and contrast them. Start by defining each term and then discussing their differences.
Answer Example
"It is common for people to confuse these terms. While they are similar, there are subtle differences between them. I would define a computer program as a set of code that executes specific tasks. This typically takes the form of an application. Computer software is any code that performs a function on a computing device. These can include the operating system, drivers, system software, APIs, applications, and other code. Computer software includes computer programs, but computer programs don't always include all the software. Computer software is more machine-centric, while computer programs are more user-centric. Is this the same way you use these terms here at Amazon?"
10. What software development process do you prefer to use?
How to Answer
This is an opening or general question an Amazon interviewer will ask to begin a conversation, get you talking, find out a little more about you, and collect information they can use for subsequent questions. They are also interested in learning whether the processes you use align with those of their organization. The best way to prepare for this type of question is to research Amazon, its software development processes, and the tools it uses. You can learn this from their website, the job description, or by speaking with former and current Amazon software development team members.
Answer Example
"The software development process I use the most is the same one I learned early in my career. It begins with a rediscovery process and requirement analysis, which leads to the specifications and software architecture. I then move on to development, testing, implementation, and documentation. After this is done and the software is fully developed, I turn my attention to training the aim users, assisting the support team to be prepared to respond to issues, and providing ongoing maintenance, bug fixes, and upgrades."
11. How would you differentiate between Quality Assurance and Quality Control as you believe it is implemented here at Amazon?
How to Answer
Anyone in the quality field, be it Software (IT), Engineering, Production, etc. should be able to differentiate between these two. Depending on what specific role you have with a task, it will be wide-ranging. In short, Quality Assurance determines if proper processes are being followed, while Quality Control deals with maintaining the quality of a software product. If the Amazon hiring manager you are meeting with has any responsibility for QA or QC in their department, they will likely ask you to differentiate the two. They may want to hear the following:
Quality Assurance - Assures that the approach and/or method used to produce a part are designed and implemented correctly.
Quality Control - A process used to test, verify, and identify a defect. It also ensures that the approaches, techniques, and methods are designed and followed correctly.
Answer Example
"Based on my research, I believe that here at Amazon, quality control and quality assurance work hand-in-hand in a software development environment. Every software development project that I have worked on required a detailed specification document that includes QA and QC as part of the (SDLC) Software Development Life Cycle.
For Quality Assurance, I attempt to prevent defects in development coding, apply statistical process controls, and define standards that need to be followed. For Quality Control, I aim to identify and improve the quality of programming code, perform validation early in the development process, and follow pre-defined processes, policies, and standards."
12. Talk about the differences between structured English and Pseudo Code?
How to Answer
Let's begin with what structured English and Pseudo Code are. Structured English is the native English language. Structured English is used to write the structure of a program module. It uses programming language keywords to achieve this. Pseudo Code, on the other hand, is like a programming language without the syntax of any specific language. This question can come in different forms, and each Amazon hiring manager will ask it differently. Pseudo Code is a kind of structured English used for describing algorithms. It describes the entire logic of an algorithm so that when implementation happens, it is translated line by line into source code. This is an algorithm question that the Amazon hiring manager will ask to see if you've written code in Lisp, FORTRAN, or C programming language. This tells the manager how you approach a project, whether it's task structured or methodology driven. There isn't a right answer, but you should ask the interviewer early in the interviewer what type of development organization he/she runs so you can answer accordingly.
Answer Example
"My approach is simple. I take an organized approach to how I write code and the way I differentiate structured English, and Pseudo Code. First, I make a list of the main tasks that need to be accomplished. Then, I break them down into small, manageable tasks that can be explained with a short phrase. It's been my experience that Pseudo Code helps me plan the structure but doesn't help me resolve the real-world programming. In my opinion, that's the bottom line difference in simple terms between the two. They both act as structured algorithms that can be used in a programming language."
13. As part of software development, were you involved with working on functional and non-functional requirements?
How to Answer
In short, functional requirements are features that are expected by users from a proposed software product. Non-functional requirements, on the other hand, are more related to security, performance, and the look and feel of the user interface (a.k.a. UI). For a project to be successful, both the customer and the Amazon project delivery teams must have a clear definition and agreement on what the project scope and detailed requirements are going to be.
There are three types of requirements: conscious, unconscious, and unidentified, or TBD, requirements.
When the Amazon hiring manager asks this question, they want to know if part of your previous jobs involved gathering and analyzing functional or non-functional requirements. For functional requirements gathering, they will want to hear how the system behaves under certain conditions within its environment, how it responds to inputs, and what outputs it generates. As far as non-functional requirements, they will want to hear about constraints on the system, quality attributes, security, performance, and availability, for example.
Answer Example
"Yes, I work with cross-functional teams within our organization on functional and non-functional requirements. I take a unique step-by-step approach to how I address these types of requirements. There are three requirement types that I use - conscious, unconscious, and unknown requirements. Conscious requirements are the basic requirements that the stakeholders are aware of. Unconscious requirements are what stakeholders would like to see but didn't communicate clearly enough to see the desired result. This stakeholder assumes that the work they envisioned will be delivered, but in most cases, it falls short of their expectation. Unknown requirements are the type that has yet to be imagined or realized. Stakeholders typically don't give input into this type because they assume it is not achievable or cannot be delivered as expected."
14. Explain how duplicates are removed from an array without using a library.
How to Answer
If we look at the core of this question, it has to do with an array, not finding duplicates. The goal here is to be able to remove duplicates from an integer array without using any collection API class libraries. There are several levels of interview questions that will come up to test your knowledge of basic to complex problem-solving solutions. This one sits somewhere in the middle of the pack. When an interviewer asks if you need a loop or recursion, (depending on your skill level) she/he is asking the order in which elements are inserted in a set. Answering with something like, 'An array is a static fixed-length structure that cannot change its length' will tell the interviewer that you have a solid understanding of how deleting an array works.
Answer Example
"Having worked at many levels using arrays and class libraries, this is a pretty straightforward answer. If your input array contains multiple duplicates, then this may result in many temporary arrays, some of which may not be needed. With this restriction in mind, I typically figure out how to minimize both memory and hardware requirements. For cases in which I need to delete an array using a more defined descriptive logic, I find duplicate elements in a given array, then run an outer loop to 0 to size. Next, I run another inner loop to find the first duplicate using another nested loop. Finally, inside the inner loop I also check for duplicate elements. If I find one, then I delete the array element."
15. Prior to starting any software development project here at Amazon, we perform a feasibility study. What is your opinion on feasibility studies, and when should it be done?
How to Answer
As a prospective Amazon Software Developer, you need to look at things from a broader perspective to gain a bigger picture of what your internal customer (department), or external customer (client) wants. In the early stages of the software development cycle, you may find that a feasibility study might be required. There are two important factors to consider before determining whether or not a feasibility study is required. A feasibility study is broken down into two important categories. The first is Technical, and the second is Operational. The Amazon interviewer wants to know if you follow protocol by offering a feasibility study to the client or just go through the motions of your job. It's important to mention this in your interview because it demonstrates that you go above and beyond the norm to get the job done right. Again, the interviewer is asking for your opinion, so it's important to express your opinion about why you think it should or shouldn't be performed along with your reasoning.
Answer Example
"In my current role, I was asked by senior management whether or not a feasibility study would provide insight into potential productivity improvements and cost-saving measures within the organization. I took on the task of planning and executing a feasibility study to determine what type of outcome we could expect. I gathered a team, formed a committee, assigned each team member a task, a goal, and a phased timeline, and asked them to report back with their findings. After the feasibility study was completed, we analyzed the data and found that there was just cause to perform a full-scale feasibility study. The results provided insightful information about what areas of the company needed further analysis and what areas had started seeing improvements. We selected the technology, production, operations, and accounting for this study. After all the data was analyzed, I provided a draft of our findings to management to review for accuracy, relevance, and completeness. I was given praise for this project since it helped the selected departments in every aspect of the operation."
16. What method do you use to find the missing number in an integer array of 1 to 100?
How to Answer
Interview questions about finding missing numbers in an integer indicate that the interviewer at Amazon has a specific type of role in mind for a candidate. More than likely, they are looking for an analytical thinker who can solve problems relatively easily. Say you have an array number from 1 to 100. These are just random numbers for now. In a sorted array, you can compare whether a number is equal to the expected next number or not. Alternatively, you can also use the BitSet method in Java to solve this problem. A BitSet solution is more general, as you can use it to find more than one missing value on an integer array. Take your time to consider how you will respond to these questions because the interviewer is testing your ability to process and come up with an answer.
Answer Example
"Having many tools at my disposal, I find that a Java solution is the best way to solve this problem. I would typically write a program to find the missing number in an array in Java, C#, or another language. Looking at this problem from a high level, this can easily be solved by calculating the sum of the series using n(n+1)/2 to solve this problem. It's a quick and efficient way to do it, but it cannot be used if the array contains more than one number or if the array contains duplicates. These solutions might require you to calculate the sum of numbers and then subtract that from the actual sum to arrive at the correct answer. Is this similar to the methodology used here at Amazon?"
17. What is software scope, and what does the process involve?
How to Answer
Software scope is a set of activities and actions that are to be performed as part of the delivery of a software product. Software scope should be well-defined with phase-by-phase milestones, functionalities, and deliverable components. A question like this will be asked during an interview with Amazon regardless of your experience level in software development. This is a fundamental question that all software developers should know. Software scope identifies a few different but important elements, including what the product will do, what is outside the scope of the project (what is not to be done), the timeline for the completion of the project, who is documenting the list of deliverables, goals, tasks, and what the estimated cost of the project will be. There's one more part that should be covered because it comes up more times than not in an Amazon interview - Project Scope creep. This is when the project experiences changes or uncontrolled growth during the project. This happens because the project was not well-defined, documented, or controlled early in the project specification phase.
Answer Example
"Software scope for me is relatively easy because, at my current company, I'm involved in every stage of the documentation and project delivery process. It is my responsibility to identify all aspects of project scope, including but not limited to what the end product will do, what the expected timeline will be, who needs to approve deliverables at every phase, what the goals and tasks will be for the assigned team members, how the features will work and their characteristics, and what the estimated cost will be for the final product. As part of my due diligence, I always draft a Project scope document that includes items that are not part of the scope of work. This helps eliminate any confusion or miscommunication with the Amazon client."
18. Do you have a preferred language that you like to write programming algorithms?
How to Answer
Every developer, no matter what level of expertise, has their views and opinions when it comes to programming language choices. It's much like a toolbox - you use them based on the specific need you have to solve a problem or fix something. One of the reasons an interviewer will ask this question is to see what range of tools you have used and which you prefer using. It always helps to start by stating that you have used several languages and are confident with all of them, but if you had to pick, choose one of your favorites. Don't be surprised if an interviewer asks why you like that language and what benefit it gives you. If the interview is going well, it wouldn't be a bad idea to talk about some of the strengths and weaknesses of a few that you have used.
As a reference, the top languages are likely Python, Java, JavaScript, C, C#, Ruby, PHP, Swift, and SQL.
Answer Example
"Every software developer has a preferred language, and I'm no different. There are a few that I like and would put on my preferred list. Let me start by giving you a few that I have found to be very useful in various environments. Perl is probably among the top for me for several reasons. Compared to Python, it's about three times as fast when reading PASTA files, and it only needs half the storage space to store sequences in memory. Java is another one of my preferred favorites, again for many reasons. Not only is it good for programming algorithms, but it can be used across multiple platforms, including web applications. Lastly, I need to include C# in this group because it requires less memory for holding strings and less time to read these types of files than most other languages."
19. If hired by Amazon, one of your tasks will be to debug yours and other developers' code. As a software developer, explain the meaning of debugging, and why it is used?
How to Answer
In the software debugging world, the process starts as soon as code is written, continues in successive stages of the development process, and is then combined with other units of programming to form a software product. Debugging is a multistep process that involves identifying a problem, isolating the source of the problem, and then correcting the problem. Please remember this since the Amazon technical manager that knows development and debugging well will quiz you on multiple facets of this area. It's important to note that hiring managers will want to determine the quality of your code. If you maintain a high-quality level of code, you'll be doing less debugging; if your code is average or not great, you'll be spending a lot more time debugging, which managers might see as a weakness. Please be aware of this.
Answer Example
"In short, debugging is an important part of determining why an operating system, application, or program is behaving abnormally. There are many things that I take into account during the debugging process. For larger lines of code, I conduct unit testing and pair programming which helps me identify bugs at an earlier stage. I also use the stand-alone debugger tool to further identify bugs. To further understand where bugs may reside, I also look at the module to see if the problem avails itself. If not, I set up a 'breakpoint' and run a program to see it run its course. If I come across errors after performing debugging and testing, I address and correct them immediately. Some examples of possible errors are...Syntax errors, Runtime errors, Logic errors, Semantic errors, etc. Does this mirror the process you use here at Amazon?"
20. What is verification and validation, and why it is important?
How to Answer
Verification and validation are extremely important in the software development process. If you can't verify or validate a set of activities to detect if the software is not implemented correctly, hasn't been built to specification, or is functioning properly, then you will surely have problems with your production environment. Amazon interviewers know how important this is and will ask questions about it to make sure that you ensure the highest quality software development. When questions come up about verification and validation, it might score you an extra point in the interview if you give the Amazon interviewer a history of how V&V was formed. It was used to design products in the manufacturing and support process areas. It's important to remember if asked that there are two important aspects of software quality management. Verification determines whether the software is being developed correctly, and validation determines whether the right software is being produced.
Answer Example
"In my opinion, verification and validation are at the heart of every development project. I take this step very seriously, and it shows in my work. This, of course, includes all the steps and procedures of validation. Prospective validation is important because it is done to ensure that the product is functioning properly. Retrospective validation is done against the written specifications and verifies actual data. Periodic validation is used to repair, relocate or dismiss data that serves no purpose. Partial validation is mostly used for research but can come in handy for pilot studies. Cross-validation is suitable for estimating the performance of statistical models. Concurrent validation is usually carried out during regular maintenance or service routines in the post-development process."
21. At a high level, describe what the software development life cycle process is.
How to Answer
SDLC, or Software Development Life Cycle, is a software development process that produces software in the most efficient way possible. SDLC includes a detailed plan for the development, alteration, maintenance, and even replacement of a software application. SDLC involves several distinct stages, which include planning, design, building, testing, and deployment. Depending on what level hiring manager you are interviewing with, you'll need to be able to answer questions related to the SDLC life cycle. If the manager is hands-on, has a history of writing code, and knows the SDLC, then you might want to mention what types of methodologies you have used. Some of the most popular ones are Waterfall, Agile, and Spiral Model. If you're interviewing with a senior-level manager who hasn't done much coding but understands the SDLC, then your answer might be at a higher level (strategic response). Either way, it's good to have a solid foundation of how SDLC works. The whole purpose of creating the SDLC foundational architecture is to lower the cost of software development while improving quality and shortening production time.
Answer Example
"I have sound knowledge of the entire SDLC process and have used it as the foundation of my development in my career. During my development career, I used a few different SDLC methodologies that I am very comfortable with. They are Waterfall, Agile, V-Model, Iterative, and Spiral Model. Not only do I fully understand SDLC, but I am also familiar with the important stages that are required for quality software development and execution. It starts with the following stages:
Requirement analysis
Technical and business specifications
Software architecture and infrastructure
Implementation/Integration
Testing
Documentation
Training and support
Maintenance"
22. Can you list the important categories of software development for me?
How to Answer
Believe it or not, the world of software development is an important part of our daily lives. Without it, we wouldn't have all the wonderful Amazon apps and mobile technologies that we use on a daily basis. Not only is software development important in our lives, but it is a highly sought-after skill for companies that can't find enough of this type of talent. Speaking of in-demand occupations, the Bureau of Labor Statistics even projected a 30% employment growth in the software development field by 2026. Let's talk about some questions that may come up in an interview. Many Amazon hiring managers like to test software developers by asking what type or category of software development they have worked on. Software developers tend to work in special development areas in which they are comfortable. Most software developers should know what the nine basic types of software development are even if they haven't worked directly in that category. For reference purposes and preparation, here are nine different types of software development:
1. Mobile Development
2. Web Development
3. Back-End Development
4. Application Development
5. Data Science Development (Data Analytics)
6. API Development
7. Security Software Development
8. Embedded Systems Development
9. Cloud Computing Development
Answer Example
"I'm quite familiar with all the categories of software development, especially those used here at Amazon, and have coded in most of them with a high level of confidence. The projects I have worked on have been Web Development - building front-end web pages in Java, WordPress, HTML, PHP, and ASP .NET, Mobile Development - building web apps on HTML5, Java, C# and Objective C, Application Development - typical app dev with tools like VB.NET, Python, C++, C#, and Java, Data Science Development (Data Analytics) - I've been building intelligent data warehouses and data sets for scientific applications using tools like MATLAB, Python, and C++, and Back-end Development - this type of work is mostly server-side database-driven development which requires different programming languages and architecture. Some of the development tools I have used here are dBase, Oracle, SQL Server, Java, Python, and C++."
23. Amazon software developers use a variety of different sorting algorithms. Tell me the difference between a Comparison and Non-Comparison Sorting Algorithm.
How to Answer
This is a common Amazon interview question asked of software developers. The question of which sorting algorithm is fastest doesn't have an easy or unambiguous answer. The speed of sorting can depend on the environment in which the sorting is done or on the type of items that are sorted and the distribution of these particular items. For example, if you are sorting a large database that cannot fit into the memory storage all at once, this would be quite different from sorting an array of 100 integers. You should also know the five Sorts that will likely be brought up in an interview. They are: Quick Sort, Insertion Sort, Shell Sort, Heap Sort, and Merge Sort. Study them and research examples of how they are used and the differences between them.
Answer Example
"One of the first things that I do before making any type of comparison is to use a test environment to test the speed of the different sorting algorithms - Comparison and Non-Comparison in this case. I test each algorithm several times for randomly generated arrays to gather the most accurate data before I proceed to the next step in the process. The next step is to look for random numbers between 0 and 10 times the array size so I can create array content. I may do a high-repetition test with numbers between 0 and 1/100 times in the same context as well. After I do this, the results may come back as completely random, sorted, or reversed. I would also run some test-cases to determine the number of repetitions from low to high for each value if the array repeats."
24. Describe how do you implement an insertion sort algorithm. What is the easiest way to do it?
How to Answer
If the Amazon interviewer asks ten software developers to describe how to implement a sort algorithm, they will probably get ten slightly different but related answers. Every software developer has a method and a work style of their own that works for them and makes them efficient at their job. The interviewer, in this case, wants to hear about the method you use and if you take any shortcuts to arrive at the final product. They want to know if you take shortcuts that could compromise the quality of work you are doing. They would like to hear about your development methodology and how you implement an insertion sort algorithm while doing your due diligence and performing quality checks along the way. In simple terms, an insertion sort starts from the index 1 (not 0), and each index starting from index 1 is like a new card that you have to place at the right position in a sorted sub-array on the left side.
Answer Example
"I can give several examples of Insertion Sorts, but I think to give you a better explanation, it would make more sense to list in detail the characteristics of an Insertion Sort. Let's start with the first one.
1. There are two types.........Selection Sort and Bubble Sort algorithms.
2. They are efficient for smaller data sets but very inefficient for larger data lists.
3. The stable sorting technique does not change the relative order of elements that are equal.
4. It doesn't take up much space. Unlike bubble sort, an insertion sort also requires additional memory space.
5. An Insertion Sort is adaptive, which means it reduces the total number of steps required for a partially sorted array to provide input."
25. Can you walk me through the meaning behind a Depth First Search Algorithm for a binary tree?
How to Answer
When looking at a question of this magnitude, there are several different factors to consider. You should clarify with the Amazon interviewer whether they want to hear the meaning for tree structure data, time complexity, extra space, or a node stack. Let's examine what some of these mean and how they might come up in the course of an interview. Time Complexity has four transversals O(n) as they visit every node exactly once. Extra Space requires O(w) Level Order Transversal where w is the maximum width of a Binary Tree that stores nodes of different levels. Similar questions might look something like, 'Tell me the typical binary tree numbers' - 1, 3, 7, 15, or in a worst-case scenario, a value of 2h is Ceil(n/2) will usually be good answers to this question. You may also be asked when extra space is required for Level Order Transversal and Depth First Transversal. Your response to that question should mention a more balanced position for Level Order Transversal and a less balanced position for Depth First Transversal.
Answer Example
"I'm familiar with the meaning of this question. There are a couple of ways to answer it, so I will give you both examples. For starters, the two most common methods of searching a graph or a tree related to a depth-first search are depth-first search and breadth-first search. Depending on the application you are using and the result you want to achieve, this should be determined by the type of data in your tree or graph data structure. For a Breadth-First Search, I start at the root node. I then search all their children nodes moving from left to right. Then I repeat the process at the level below the root node. I typically repeat this on each level until I reach the end of the tree or node. As a general practice, I use a queue as an intermediary step as a way of keeping track of what nodes I need to search."
26. What is your experience with implementing a Binary Search Algorithm without recursion?
How to Answer
This is the type of question an interviewer will ask if he starts to doubt your ability to perform certain duties revolving around algorithm-based development. Since software development at some level relies on algorithms and data structures, this will be an important question that you must understand and answer correctly. When the binary search structure is broken down, it contains a binary search or half-interval search - a divide-and-conquer algorithm that seeks a position of an item in a sorted array. When a hiring manager asks about comparing inputs and outputs to the middle element of an array, they are asking if a search returns the position of an element. It's essential to know and understand this concept. Another question that may be asked is whether or not an input is less than or greater than an element. If you have a sound understanding of these concepts, ask the interviewer what level of detail they would like in your answer. Some interviewers only want to hear a high-level answer, and others want a detailed explanation.
Answer Example
"I have extensive experience and hands-on knowledge implementing binary search algorithms with and without recursion. For this answer, I will talk about BSA without recursion. Binary search algorithms typically halve the number of items checked for each successive iteration, thus locating the given item in logarithmic time. Furthermore, a Binary Search Implementation in a Java algorithm is implemented recursively respectively. Also, I would like to share the interesting fact that binary search algorithm implementation is mostly done without recursion. It is also known as iterative binary search."
27. How do you approach implementing an LRU Cache in your favorite programming language?
How to Answer
Having a favorite programming language reveals a lot about the experience and expertise level of a software developer. There are numerous languages to pick from, and depending on what applications you are working on and what you intend to achieve, some might work better than others. The Amazon hiring manager might start with a question about your favorite language and then move to your approach and possibly your methodology or reasoning for favoring that language. If we intend to use Java for this example, we would look at the LRU cache and recently used entries, then remove the element from the bottom and add an element to the start of a LinkedList; wherever any entry is accessed, it is moved to the top so that recently used entries will reside at the top, and the least used entries will reside on the bottom.
Answer Example
"The way I approach implementing an LRU Cache is based on what programming language is used in the development environment. Yes, I do have a preference, but I consider it a tool in a toolbox. I use whatever is available to me to easily adapt my process to the ones used here at Amazon. Java seems to be a tool of choice in most development environments and for me as well. Say, for example, we are given total possible page numbers that can be referred. We are also given cache (or memory) size (number of page frames that the cache can hold at a given time). The LRU caching scheme is intended to remove the least recently used frame when the cache is full, and thus a new page is referenced which was not in the cache before. Examples of this approach and method are referenced in the Galvin book."
28. Based on your experience, what's the best way to find a node that begins with two single link lists?
How to Answer
You will find that Amazon interviewers like to ask candidates about the various methods they use to find nodes and link lists. Don't be surprised if this question comes up a couple of times in an interview, but is phrased differently. A linked list is easiest explained as a linear data structure with the collection of multiple nodes where each element stores its data and a pointer to the location of the next element. The last link is essentially a linked list that points to null, which indicates it's at the end of a chain. On the other hand, an element in a linked list is called a node. The first node is called the head, and the last is called the tail. An interviewer will likely dig into questions about linear data structure and which nodes contain a value and pointer.
Answer Example
"The best example I can give is to list the most important properties of a class Linked List. This highlights how they are used and why they are used.
A Linked List maintains an insertion order of the elements.
Implements Queue and Deque interfaces can also be used as a Queue, Deque, or Stack.
A Linked List can contain all the elements, including but not limited to duplicates and null.
A Java Linked List Class library is not synchronized, which means in a multi-threaded environment, it must be synchronized concurrently for external modifications to the linked list.
A Linked List Class doesn't implement a Random Access interface so that elements can be accessed in sequential order.
I can use a List Iterator to iterate elements of the list.
I can use a collections synchronized List (new Linked List) to get a synchronized linked list."
29. How do you find the starting node of a cycle if a link contains a cycle?
How to Answer
Let's say for argument's sake that two cyclists are pointed at the beginning of the starting line with their bicycles. For clarification purposes, the two cyclists are the pointers. In this example, if we move the first cyclist one step at a time and the second cyclist two steps at a time, they would eventually meet at a single point. An interviewer wants to know whether you can explain how you arrive at an answer and the reasoning behind your findings. This is important because when you explain if a link contains a cycle, you need to back that up with a verbal or written (whiteboard) example. A typical response might be to mention that a meeting point is 'P' steps away from the beginning of the cycle, and the 'cyclists' meet when cyclist one has taken a '7' total steps toward the point.
Answer Example
"The way I find a starting node of a cycle is to measure the length of 'R' and how far the distance is from the cycle. That's a simple explanation. A more complex illustration to explain this depicts two cyclists. When cyclist 1 travels double the speed of cyclist 2, they both arrive at the same meeting point when time is constant. Since there are several examples I can give to make my point, I would like to share one more. The next pointer of each node that is traversed is made to point to this temporary node. This way, we are using the next pointer of a node as a flag to indicate whether the node has been traversed or not. If I come across a node that points to null, then the loop doesn't exist. If I find that the code runs in O(n) time complexity and uses constant memory space, then that tells me that this needs to be researched further."
30. Here at Amazon, our developers work with Java quite a bit. Tell me how you find duplicate numbers in an array in Java containing multiple duplicates?
How to Answer
You will find that Java is used with many applications, on many different platforms, and in a multitude of languages as well. If you're in software development and haven't learned to program in Java, you should learn it as soon as possible. It is becoming a core technology for web development and development in general. It has many capabilities that other programming tools don't have, and that's why it's been a favorite for many developers and companies such as Amazon. Everyone in IT has their way of approaching a complex problem and finding a solution. In this case, we're talking about finding duplicate numbers in an array in Java. Luckily, Java is a go-to for most developers trying to solve this problem. An interviewer might phrase this question differently. When you hear the words Sorted or Binary search, they are looking for you to explain what inner and outer loops mean or parsing items inside an array. Be prepared for those questions as well.
Answer Example
"Problem-solving has always been a strength of mine. I enjoy the challenge of finding a solution to a problem in the shortest time possible. I was working on a Java development project that required using Binary Search and Sorting in a recent project. I started by adding a duplicate list of elements inserted back into an array. I also used two pointers to solve this, comparing an element and its closest neighbor to maintain a distinct element. Another method that I tried was parsing all the items inside of an array containing 'n' to give me O(n). This way, we get an array of the n+1 element with integers between 1 and 4, which means there will be at least one duplicate as a result."