Amazon, which was originally an online bookstore, has become an ecommerce giant that also develops cloud computing, digital streaming, and artificial intelligence.
No wonder Amazon needs a lot of machine learning engineers. In fact, thanks to them, the company can sell more than 12 million products to more than 100 million Amazon Prime subscribers and deliver them in one to two days. Let’s figure out what such a specialist in Amazon does, what skills he should have and what should be expected from an interview.
How to Become a Machine Learning Engineer at Amazon
What Amazon Machine Learning Engineer Does
Amazon's machine learning engineer processes large datasets to create machine learning and deep learning models for various company customers. This specialist will work with consultants from Amazon Professional Services, lead projects from start to finish, and assist in putting the resulting models into practice.
What skills should he have
- Bachelor's degree in mathematics, statistics, computer science or other related field
- From five years of work in the relevant field
- Experience with a wide range of forecasting and decision-making models, as well as with data mining methods and tools for developing such models
- Experience in creating and operating highly accessible distributed systems for extracting, receiving and processing large data sets
- Experience using Linux / UNIX to process large datasets
- Experience with AWS technologies such as Redshift, S3, EC2, Data Pipeline and EMR
- A combination of technical and business skills sufficient to interact with all levels and directions in the client’s company
What positions can he count on
Amazon has several positions for machine learning and data science professionals. These are data scientists, machine learning engineers, researchers of fundamental and applied problems. Although they are similar in nature, they still have significant differences.
- Data Scientists pay special attention to data analysis and are the link between business and the technical side of the matter. They are responsible for analyzing and modeling large data sets.
- Machine learning engineers – These are experts in building machine learning and deep learning models. They create models not only for Amazon, but also for other large companies. These specialists are also responsible for the implementation of the models and their preparation for work.
- Researchers of fundamental problems usually have a master's or doctoral degree. They are supposed to go beyond the scope of the possible. These scientists test Amazon’s old and new technologies to determine how useful they are in practice.
- Applied researchers also usually have a degree. At Amazon, their role is slightly more significant than that of researchers at fundamental problems. These specialists focus on projects that are aimed at improving the quality of service: automatic speech recognition, natural language recognition, audio processing, text to speech, dialogue management, and so on.
How is the interview
The first telephone call is made by a recruiter or hiring manager: this conversation is needed in order to get to know each other better. As a rule, the interviewer briefly talks about the position and the interview process, and also asks standard questions about the resume and work experience. He needs to make sure that the applicant is interested in working with Amazon and has the necessary skills.
It is usually conducted by the machine learning team manager. First, he will ask a number of general questions about the fundamental concepts of machine learning, for example, he will ask to explain the difference between the types of models or to talk about the dilemma of bias-dispersion and retraining.
Next, you will need to write code in any convenient programming language. Examples of Amazon tasks and their solutions can be viewed here.
It consists of five to six meetings, at which the candidate for the position is asked questions related to both the technical aspect of the work and the behavioral one.
- Technical issues. You should expect at least two conversations on the concepts of machine learning and programming. It is known that job seekers can even ask questions about object-oriented design.
- Behavioral issues. They can be heard at all meetings and will cover aspects such as professional experience, the reason for the last dismissal, and attitude to teamwork. In addition, a job applicant will almost certainly be asked about Amazon’s leadership principles. There are a total of 14 pieces, you can familiarize yourself with them at website company.