Amazon Leadership Principle Question

can you recognize strong performers?

- In the beginning, talk less and listening carefully every one opinion - more likely to perform their own research and spend more time on projects. giving evidence from their research to prove their opinion instead of an assumption. - cool under pressure, easily adapt to new things - Do over expectation - Proactive

How long do you expect to work for this company?

As long as we are both satisfy with each other. I don't have any immediate plans to move on. I'd like a job where I can have continuity and be part of a team. And [company] is one of the top company that I want to work for.

Why should we hire you?

Comparing my background as a business data analyst and job description for this role. I feel I have the skills, qualities, and attributes to perform the job to a high standard. I have 6 years of using advanced SQL and Excel including 2 years of professional experience. 2 years of using Python, R, Tableau, Power BI in an academic environment. My academic performance can prove my competence in these skills. I am very adaptive, flexible, resilient person. I have a good work ethic, a willingness to learn new thing which help me to develop myself into new environment and problem. For example, 2 years ago I had come to Seattle to pursue my master's degree, this is the most difficult time for me. I live in a new country, with new language, new culture, and customs. But I had adapted quickly to strengthen myself and accelerate my career. Finally, Amazon is the company that I most want to work for. I learned about Amazon a lot when I was in Vietnam. The stories of Amazon gave me much inspiration. I think amazon is place that I will have chance to develop my career as well as make impactful contribution. I think I will be a good investment in the long-term.

What do you think are the most important quality for someone in this role

Critical thinking an problem solving is the first important quality. Working as a data analyst we need dive deep into data and interpret them. With a critical think and problem solving in mind, that can help the analyst make an correct question from then they can derive the answer. Second is an effective communication and presentation that can help the analyst interpret their analysis to the stockholder and deliver the result correctly. Finally, I think that should be an passion about data manipulation and data-drive decision making. Data analysis mean that we have to work with data, number and sometime it's boring for people to look at a large number of data, column without an understand of what their meaning. So a data analyst must have a passion and interest in working with data. Typically to learn a new skill set I would prefer to follow an official course It could be online or face to face cource. I would like to gather some basic idea about the skill before I jumping to self study and advance topic. The when I became familiar with fundamental knowledge I would like to pic some Project

How do you find the time to stay inspired, acquire new knowledge, innovate in your work?

I find out the satisfaction of making awesome things from nothing, including learning new things. one of my favorite activities in my work that builds some tools to automate my tedious and repetitive tasks. I learn that from my colleague. for example, I had built a SQL database that can automatedly implement ETL, Integration, aggregation, and sent a report to my boss every day. It cut out the time I spend on running analysis manually.

Tell me a time when you could have stopped working on something but you persisted.

I finished my independent study project a couple of weeks ago. What I studied is applying machine learning to classifying fake reviews. In this project, I tried to use NLP, Text Mining, and Machine Learning algorithms to analyze the review content. This is the first time I learn text mining and NLP. The research yield very satisfactory results. but there are some limitations that I think I can do better if I have more time and resources. Due to time constraints, I cant do more experiment such as combining different data sources, deep learning, tried some different setup. I have to finish and submit to my professor. However, I will come back to the project in the future after having a job and get over these days. I have plenty of ideas to solve problems.

Have you ever had difficulty working with a manager

I have been very lucky to have had terrific managers during my career so far. I have respected each of them, and got along well with all of them.

Tell me about a time you made a serious mistake at work.

I received a support request to transform some loans from brand A to brand B. At that time, we did not have an official procedure to handle that kind of request. I had consulted to these branch some options and ask them to create an help-desk request. Unfortunately, I forgot to ask the help desk team to assign this request to me when it came up. Then It was assign to my colleague and he did it wrongly that made the balance sheet went wrong. I then immediately, trace back to his code. There was a hundred loans that was transform incorrectly. I had stay late that night to fix the problem

Overcome the opstacle:

I started to analyze the inventory data to see how paccar manage their inventory, I applied basic statistical analysis to get an overview and point out some outliers which were storing to much but low turnover and storing too less but high demand. I also noticed when paccart restocked their inventory. From then, I developed 2 metric which presented paccar inventory management method which is time to survive = number of days that paccar can meet the market demand before they out of stock and time to recover = number of days their inventory will be restocked. By using these 2 metric we can point out which product were overstock and understock. The other number that we needed to consolidate our analysis is the daily future demand. To get this number, I had applied machine learning to predict the future demand based on the history sales data. After that, I can form a threshold line to help me classify understock and overstock product. Finally I visualized my analysis into a scatter plot

What is your biggest weaknesses

I think I have many weaknesses, but one thing that makes me think a lot right now is public speaking. I usually get nervous when speaking in front of many people maybe just more than 20 people can make stumble over a word. The reason is that English is not my first language so I easily forget the thing that I want to say. One thing that I did to overcome this obstacle is attending a public speaking training organization named toastmaster. I also try to get out of my comfort zone and take a chance to speak with people. sometimes it's really embarrassing. But I think it helps.

Tell me about a time when you had to work with limited time or resources.

I was assigned for a task that generates the system performance reports. I figured out that the T24 system that we used to operate the bank business already had a module that records every patch job of the system. I suggest we used this module. Unfortunately, the module is not a built-in function, we have to pay to use the reports, and it cost a lot. At that time, the bank did not have the plan to upgrade this function. So I have to create a report manually from the raw data. After a few weeks, I found that it was possible to create a tool to automate this task, and I can use the resource I already have to do it. I built a script to extract, transform and load the data from the T24 system to SQL server, then I create Store Procedure and Patch job to analyze and send the report to my manager. My solution save the time from 2~3 hours every day to zero, my supervisor had an in-time report every day, and it only cost me a few weeks to do.

Tell me about the most difficult interaction you had at work

I work in a group of 3 for database architecture class. one team member usually procrastinated his duties and that frustrated other members. a day before the final submission, he had argued with other teammate jobs because he thought she could do better if follow his idea. I and She both disagree with him because it was the last minute and her work is also very good. He has more than a week to suggest his ideas but he told at the night before the due date and when we still need to prepare for the presentation. the argument did not stop. So I said, we can not continue arguing like this we must finish it tonight. I will cover her jobs. because she has kid to take care of. I and he worked until 2 am to complete the project. eventually, we got A for this class, we then gave him some feedback and advice for his procrastination. He agrees and apologized to her.

Give me an example of when you took a risk and failed.

I worked in a group of 5 people in boardroom analytics class to conduct Sale and Marketing Attribution analysis for Zillow group. That time, we only have 2 days left before we present a solution to guess speakers. we had very long meeting that night but can not decide which solution we should present. No one wants to be responsible for their decision. I suggested choosing the one that I most understand and said 'if we have new information we can come back to the decision in tomorrow'. every one agreed. Unfortunately, our solution did not meet the expectation from guess speaker, we got B for that project. I apologized for my mistake. The lesson I learned was that we should not make decision with insufficient information. I came up 3 strategy to get more information. 1- Ask more space:Contract the stakeholders for more information and see if more time is available to make choice. 2- Compare options: never jump to the first choice unless I familiar with all of others. 3- Check whether I am biased. In this case, I was biased

Tell me about a time when you had to deal with a poor performer on your team.

I worked in a group with 3 people in the database architecture class. A teammate usually procrastinated his task until the last minutes and complaint about other teammate's tasks. My team stresses out because it took time to consider his feedback and hard to implement the changes. At the short-term measurement, I had to change team dynamics. I suggested to other teammates that we will cover his tasks and prepare it well in advance as a backup. we set the deadline earlier and push him to follow our schedule. Although It creates more work for us, it gave a great chance to learn more than just our part. We got A for this class. We then gave him some feedback , and point out how he should change in the future. Latter I had to work with this guy again, I set up some rule with team such as prepare before the meeting, finish everything one day before the death line, and we won't change the plan in the last minute unless it is necessary. That helped we guide the team.

Tell me about a time you disagreed with a colleague. What is the process you used to work it out?

I worked on a group project with team of 5 people. our project is to analyze customer behavior and predict their services adoption. I and one teammate had argue about the ML algorithm that we should used. Due to the earlier experiment, I saw that SVM is not a good fit in this case, I suggested to move on with other algorithms like Random Forest or Neural Network. He did not agree with my opinion and want to work on SVM because He used it in his internship and it worked well. I had show him the result from my experiment why it was not good. But he still persistent to his idea. I suggested we will do both of them and I let him responsible for SVM. In the end, when we come back for decision. He agreed with me that SVM was not a good choice.

Describe how you would handle a busy situation where three people are waiting for help from you.

I would maintain a positive outlook and utilize my organizational skills to develop a plan of action.. Typically, I would ask and listen to their problems carefully one by one first. Then I will evaluate these problems by the level of complication, importance and time-consuming even the problems are unachievable. If there were any tasks I could delegate to other capable individuals, I would not be afraid to ask for assistance.

What are your strengths?

I would say that I'm high self-discipline and a well-organized person. I have applied my natural skill for organizing tasks and projects to all aspects of my life. One secret to being successful that I learn is being ready and prepared in the early state. For example, I'm doing an independent study on Fake review this quarter. To obtain the best result from the study I had prepared it for 2 months in advance. Because it requires me knowledge in NLP, text mining which I never learn before. It also takes time for me to get the datasets, also read all related papers.

Where do you want to be in five years?

I'm really excited by this position at [company] because, in the next five years, I'd like to be seen as someone with deep expertise in the Data Science, and I know that's something that I'll have an opportunity to do here. I'm also really excited to take on more managerial responsibilities in the next few years and potentially even take the lead on some Data Science projects. I've been lucky enough to work with some amazing managers, and so developing into a great manager myself is something I'm really excited about.

Why are you interested in the job?

My long term goal is becoming a data scientist. To do it, I have to gain knowledge and experience in several field such as data analytics, data engineering, and Business intelligent. The first is job is very important to my career. Comparing this job description, my goal and background. I feel I have the skills, qualities and attributes to carry out the job to high standard. Because, it required the skill that I most proud of SQL, in addition manipulating data to extract an insight which support for decision making is what I passionate about. All of them reflect the reason why I want to become an amazonian

Walk us through a time when you helped a customer through a difficult process and what that looked like.

In my previous job, I did not work directly to customers but indirectly. I have to support the teller to solve any problem related to t24 system. there was a time that I received a call from teller during the break time, she had trouble when committing a transaction and the customer was waiting for her. It wasn't a serious problem, but if I follow the bank policy it would take 20~30 minutes to solve the problems. I know that I cant not let the customer waiting until the office back to work. Fortunately, my account still has the authority to access the system so I kept heading roadblocks, tell my boss what I was going to do, and will fill in the process latter. then I did, the problem was solved, the transaction completed, the customer can back to work, the teller wasn't marked as low performance.

Explain and give an example of your routine quality checking habits or processes

In my previous role, I usually had a lot of major projects going on and I was responsible for my day to day duties as the same times. Whenever I received a new assignment I always list all tasks that need to be done to complete the assigment, I put all of the task in the same pool named waiting task. In order to get all the work done, I had to take a step back to assess which task took priority over others base on their difficulty, time consuming and death line. At the beginning of the day I pick top 3 of them, and work down from the most important to not as important . I will try to finish it as soon as I can, the pick another three. If I need help I would be sure to ask my colleagues . For example, when I developed new module for microfinance, I have list all of Screens, class, table, Store procedure, reports , ect. That I need to modify or develop. By following that way, I won't miss any single tasks, and I also can add more to the list if it's necessary. It help me organize my task, and every day I don't waste time on planning what I am going to do for that day.

Describe your career goals

In the short term, I hope to work as a data engineer for a company such as yours; one with a mission-based on handling data and analyzing. Working as a [position] for your company I believe in will prepare me to hopefully take on expanded responsibilities in the future such as Data Scientist. As a data scientist, I need to deeply understand tools and concepts from several fields including data engineer.

Tell us what has been your best work achievement to this point in your career. What made this achievement important?

Independent study: fake review. Applied machine learning and text mining technique to classify customer review whether it's truthful and untruthful review. This is first research that was conduct under professional environment with the guidance from a preeminent professor at Seattle University.During the research I have face multiple obstacles such as finding reliable data source, learning text-mining, handle large scale data set with high dimensionality, apply big data analytics technique. I had overcome these obstacle and complete my paper. My research will make a great contribution in solving deceptive customer review online which effect on both merchant and consume

With regards to a work project, please share whether you prefer to dig into the project right away on your own to get familiar or become comfortable by being directed and guided.

It's depend on the tasks I will apply either approach. I would prefer to dig into the project right away and do my own research to get familiar with the problem. In this way, I can gain my experience in problem solving and troubleshooting as well as leadership. I can get more information about the project in this way, I can help me for learning project planning and management which are useful for the future project. However, If the task already have some sort of solution, I would like to listen the guidance and direction from other first, then digging into the project deeper. In that way, I will able to narrow down the scope of solution, save time, and by combining with my research I can probably deliver a better solution based on the existing one.

Describe your work style

My work style is extremely flexible. After working on so many different projects and people, I realize that I need to be adaptive. I can handle multitasking, but I prefer to work on one project at a time, working as quickly but efficiently as soon as possible. That's Why I always start my project immediately as I can. I don't like to procrastinate my task

Dive Deep

One of my favorite topic in data science is Natural Language Processing. During the time at Seattle University, I had a chance working on a project to detect spam customer review on social media. It's very interesting topic also challenge. Because, I never learn about NLP before also deep learning which are two important technique to analyze text data. I started analyze the data by learn how to clean and tokenize the data. I then Isolated the review that were marked as fake review for analyzing. I tried using Text mining and Machine learning techniques To indentify the different patterns between real and fake review. First I tried bag of words, n-gram, tf-Idf and tried to cluster them into group to see the difference. I had read many different paper for reference, and it was so surprising that my result did not show any useful information. I had accepted that by analyzing only the review content we can not differentiate real and fake review. Because over years, fake review maker also tried mimic the real review to improve. I got stuck, I tried work around the problem. I realized that even though the review content can not help to classify the review itself, but if I combined it with reviewer data- who wrote that review, the business data that review aim to I can cluster the reviewer into different group and it can help me to classify fake review. It tooke me a very deep dive in the data to find a useful set of attribute that can be used in detect fake review most effectively. At the end of the research, I came up with a conclusion that instead of trying to analyzing the review content, we should focus more on analyzing the reviewer which yield a better way for classifying.My research was not ended yet, It's just a beginning for another research which can help us prevent the effect of fake review on online market.

Tell me a time when you earned the trust of a group.

One thing that I always keep in my mind that "Never say that's not my jobs". I was always voluntarily to a new project no matter whether these projects belong to my responsibility or not. If they need a person and I am able to do that I will join it. For example, in the first months of my career, my boss needs a tool to help him listing and organizing all the functions appear in the system, where and what they used for. Nobody wanted to do it at that time, as a newcomer I had a lot of free time that i used for learning and training. I thought that was a great opportunity for me to practice what I had learned. So I moved ahead and take the chance. to be honest, this task was really hard. But with the support of my colleague, I successfully completed it after a few weeks. I am not afraid to admit my weakness and asked for assistance. From then, my boss and my leader acknowledge me as a strong performer and reliable person.

How do you develop client relationships?

Sincerity is the key factor. I always remind that when I talk to tellers. I tried to remember their name, their company. try to be their friend who are helping them solve their problem. and after helped them address the problem, I usually call back to them, make sure everything works smoothly.

Describe a difficult task you were faced with and how you addressed I

The situation is when I worked on my independent study project about Fake Review Detection The tasks was: Professor wanted me to apply unsuppervised learning to segment the review dataset. He suggested I cluster the dataset with EM clustering with 10 cluster. The problem was that due to the large scale of processed text-data, It was high dimensional data set. I can not compute my model in my laptop, It's require a more powerful computer. The action came out is: I tried to apply skill and knowledge that I learn from big data analytics course. First I had applied Spark, PySpark to analyze the data. This solution did not help much, because It was still require a lot of computation due to high dimensionality. After all, I still need a better computing power. So I decided to move my code to cloud. I create an AWS E2C instant with 32 GB of memory to compute my analysis. The result was that with AWS E2C I can easily execute my code. It was not only help me to solve the problem, but also enhance the entire of my analysis process and gave me more time to interpret the result. I got A for this project.

dive deep

The situation was when I work in paccar inventory optimization project. our mission was helping paccar to find out the problem in their inventory which cost them money and accelerate their distributing process. There were different kind of product that was stored in their inventory which impacted on their inventory management decision such as where and how long to order products, sized, history sales, predicted sales, seasonality can also be a factor. I started to analyze the inventory data to see how paccar manage their inventory, I applied basic statistical analysis to get an overview and point out some outliers which were storing too much but low turnover and storing too less but high demand. I also noticed when paccar restocked their inventory. From then, I developed 2 metrics which assess paccar inventory management method which were time to survive = number of days that paccar can meet the market demand before they out of stock and time to recover = number of days their inventory will be restocked. By using these 2 metrics we can point out which product was overstock and understock. The other number that we needed to consolidate our analysis is the predicted daily demand. To get this number, I had applied machine learning to predict future demand based on historical sales data. After that, I can form a threshold line to help me classify understock and overstock products. if the product appeared on that line mean it's TTS=TTR, above mean overstock and beloved mean understock. To optimize paccar inventory we need to move all product above the threshold line but not too far. I suggested a confident interval with 10% this number might be different to specific product depend on paccar. Finally I visualized my analysis into a scatter plot to demonstrate my analysis. The result was that our recommendation can help paccar save 2.x million of dollar. We got many positive feedbacks from paccar teams.

Give me an example of when you did more than what was required in any job experience.

The situation was: when I worked in my independent study class about Fake customer review detection.

Tell me about a time you used customer feedback to change the way you worked.

There was a time a Teller gave feedback to my manager about me. I had a meeting with my manager and figure out the solution with him. The problem was that I handle a conversation too fast that make some teller can not follow my instructions. I sometimes understand the problem before they complete explain it to me, and to save time I immediately jump into the solution. I recognized it some time and slow it down when the teller told me. My manager suggest a solution for me. While I talking on the phone, I must write it down on paper and frequently asks the teller whether they get what I said. These steps will slow myself and make sure the teller will understand my instruction. and It work, The result was that I got some positive feedback from the teller a few months later. I also apply this technique when I discuss with my colleague or during meeting.

Tell me about a time when you had to dive deep into data and the results you achieved.

There was a time that a function that I developed went wrong after a year of migration. the error was raised when I had a meeting. The problem was really annoyed me because I heard all the complain right after I left the meeting room. So I immediately open my code and read carefully everything relate to it and realize that a part of my code was modified without notice. Then I trace back to migration history and see the logs. My code was accidentally changed three days before by my colleague. I fixed it, I also tune the code to reduce the complexity. He apologized me, It was not a big problems, I just reminded he writing some note to his code for next time.

"Tell me about a time you needed information from someone who wasn't responsive. What did you do?

This question remind a time when I was working on a group project in Econometric Class with a group of three. Before submitting our final paper, we had a meeting with professor, so we had to complete our first draft before the meetings. I had emphasize with my teammate we had to work through the weekend in order to complete the draft on Sunday, then we would have Monday to review everything before the meeting on Tuesday. However, during that weekend we cannot contact with one of team member and he did not send any notice to us before that day. What I did was first to discuss with other teammate to consider every option we had. We had review all the work, divided the work of the missing member to all others. Next we sent him an email which explain everything we did to cover his stuff. Fortunately, we finished the draft in time before the meeting.

Have you gotten angry at work?

To be honest, I have, It's usual that something out of your control and messes up everything you are doing. So do I. I'm angry but I do not lose control. When I get stressed, I'm quiet, step back, walk around, get some fresh air, thoughtfully think through the situation and then begin to formulate a plan of action.

Why Amazon?

Well the Amazon reputation is certainly a first factor. I would be proud to work for such an innovative company. I heard and read a lot news about amazon from I was in Vietnam. The Amazon stories gave me a lot of inspiration about entrepreneurship, leadership, and start up. The second reason is that AWS is a leader in cloud computing services, also in data science field. That means there a lot of challenges as well as opportunities for me to apply my skills and make contribution. .I consider getting a job at amazon is a big goal to challenge myself.

Do you take work home with you

When I need to, bringing work home with me is not a problem. I realize the importance of meeting deadlines and getting work done on time, and sometimes that requires extra hours in the office or at home.

Tell me about a time you handled a difficult customer.

a leader of the sales team needs support for a loan. there was a problem and he can not submit the loan to Disburse money to the customer. He was angry because the help-desks had put he through multiple time but can not solve the problem. I then took the call, I listen to him carefully to find out what the problem was. I then acknowledge his frustration and complaint. after a few minutes talking to him and dive deep into data and system logs. I figure out his problem and what was wrong with the IT-helpdesk because he described the problem incorrectly that made IT help-desk misunderstood and put him through a wrong person. Then I talk to the team that responsible this issues and explained it carefully, and switch him one last time to the correct team. I also instruct him on how to explain the problem correctly for the next time. He then sounded calmer and thank for my help.

Tell me about a time when you sacrifice short-term gain for long term-goal.

the situation was in my previous job. When I developed a new micro-finance module. The business department required me to finish it in 3 months. However, based on my previous experience working with them, I thought it was really better to build a flexible, adaptive, and expandable module. By troubleshooting for problems that didn't appear, I can prepare for future issues. Although it can cost more time at that time however, it will reduce a lot of work when we need an upgrade or modification in the future. The result was that after 6 months, the central bank changed its policy for microfinance and we indeed need to upgrade this module. That time, I and my team only need few weeks to modify and deploy a new upgrade because everything was prepared already.