What is an AI engineer whose demand is increasing due to overheating AI development?

Recent AI development such as “5G”, VR (Virtual Reality), cashless, etc., which are attracting a lot of attention for automatic driving of cars and image recognition AI, are developing at a tremendous pace, and AI engineers are It is needed all over the world.

In this article, I will explain the points to become an AI engineer, including the outline and current status of AI engineers.

table of contents

1. What is an AI engineer?

AI (Artificial Intelligence) means “computers perform the work of human intelligence → artificial intelligence “. AI engineer is literally a general term for engineers who specialize in the AI ​​field.

1.1 Types of AI engineers

What kind of engineer is an AI engineer?
The word “AI” has a wide meaning, and people with various roles are involved in the development of AI technology. Among them, AI engineers often refer to machine learning engineers and data scientists.

1.1.1 Machine learning engineer

Machine learning engineers aim to give AI the data it needs, make it smarter, and connect it to problem solving.
Making full use of knowledge of machine learning and deep learning, as “pre-processing”, we combine scattered data to improve the quality of data input to AI, and build and implement machine learning models.

◆ Machine Learning Allowing
machines to make their own decisions by giving them the ability to learn

◆ Deep learning Deep learning
contained in machine learning is recognized by the machines themselves. · To be able to understand and anticipate new things

1.1.2 Data Scientist

A data scientist is an expert who solves various problems by extracting necessary data from big data with a huge amount of information and performing statistical analysis.
It also has a strong consulting element, and has the role of producing AI models and analysis results that are likely to be useful for improving corporate operations.
We analyze the developed machine learning algorithms and often receive consultations on machine learning projects, expanding the scope of our work.

1.2 What is different from an IT engineer?

What is the difference between an AI engineer and an IT engineer?

◆ Differences in work content ・ We will develop the system according to the specifications and plans created based on the
IT engineer ‘s requirements ・ AI engineer Understand the nature of AI, give data that AI can easily learn, and analyze problems We will devise and develop solutions to problems by utilizing AI.

◆ Differences in required skills
・ IT engineers
Mainly knowledge about information systems, programming knowledge
・ AI engineers
Programming knowledge (Python, etc.), mathematical knowledge, statistics, deep learning, knowledge about machine learning

2. Current status of domestic AI engineers due to expansion of AI technology

2.1 Japan’s AI field lags behind the rest of the world

“Japan has become an AI underdeveloped country before I knew it. Until recently, Japan was the most advanced country in the world in terms of technology, but it is perfect in the field of AI, which has been the most innovative in the last few years. It has become a developing country. (Omitted) There is nothing I can’t recover yet, but it’s pretty bad. ”
” Because I grew up in Japan and loved it, I can catch up with and overtake other countries (in the field of AI). I think I have to

(Source: SoftBank World 2019 / President Masayoshi Son points out in the AI ​​field )

The world’s AI development competition is centered on the United States and China, and AI entrepreneurs are being born one after another from India and Southeast Asia.

In addition, at the international conference ” NIPS 2017 “, which is said to be the highest peak in AI-related technology, companies and universities with a large number of papers in the AI ​​field are said to be mainly in the United States, followed by Europeans and Chinese such as the United Kingdom and Switzerland. It was a composition.
Only the University of Tokyo and RIKEN are in the top 40 in terms of the number of accepted papers. There were no private companies.

From these things, it can be said that Japan is lagging behind.

2.2 AI engineers have a serious shortage of human resources

◆ Acquisition of AI human resources is a battle
for AI Human resources are needed to advance AI development, but the supply is overwhelmingly insufficient.
Each company is actively focusing on acquiring AI engineers both in Japan and overseas.

[Examples of initiatives by each company]

・ Major companies such as Panasonic, Yahoo, and NEC are focusing on securing AI human resources and strengthening in-house training.

・ Joint company information sessions for foreign engineers in India, China, North America, etc. were held, and each company focused on acquiring AI engineers.

・ Large companies such as Mercari and Yahoo are participating in securing human resources for elite engineers of Indian Institute of Technology, the highest university in India.

2.3 AI development and AI engineers concentrated in the Tokyo area

◆ There is an AI test, 70% of which is an AI online qualification test G test (generalist) conducted by the
Japan Deep Learning Association in the Tokyo area . The relatively high pass rate of over 50% does not seem to be a level that beginners can easily pass, as the test takers are mainly those who already have skills. * A generalist is a “human resource who has basic knowledge of deep learning and has the ability to determine an appropriate utilization policy and apply it to business.”

More than 70% of the G-test test takers are concentrated in the Tokyo area, including the three prefectures of Saitama, Chiba, and Kanagawa, and there are few test takers in rural areas, so there is a clear gap in interest.

There are many AI-related projects in the Tokyo metropolitan area, and many AI engineers work in the Tokyo area.
In the Tokyo metropolitan area, where interest in the AI ​​field is extremely high, there are still few places to utilize AI in rural areas, and it is possible that people will not feel familiar with it.

2.4 Reasons why AI engineers have high annual income

There are three main reasons why AI engineers have a high annual income.
・ Shortage of AI human resources
・ High expertise
・ Highly evaluated worldwide

In addition to the overwhelming shortage of human resources mentioned in the previous section, AI engineers are a highly specialized field.
Currently, most AI engineers are mainly masters and doctoral degree holders who have studied at graduate schools in information science, and if they do not have specialized knowledge, they will learn a language such as Python and use machine learning. There are many patterns of being involved in such things as becoming an AI engineer.
The progress of AI development is remarkable all over the world, and talented AI engineers with advanced skills have very high annual income.

3. How to become an AI engineer?

An AI engineer who is attracting attention because it is expected that demand will increase in the future and high income can be expected. How can I be?

3.1 Essential skills for AI engineers

◆ Knowledge of programming skills and libraries
Mainly Python.
Python has a library of useful for machine learning. Knowledge is required to master various libraries such as Tensorflow and scikit-learn, including a library that is convenient for data collection and analysis – Pandas –.

◆ Mathematics knowledge
High school mathematics level knowledge is essential. Calculus, statistics, linear algebra, matrix mental arithmetic, etc.

◆ Machine learning knowledge

◆ Knowledge of database operations including SQL grammar

◆ Cloud knowledge
Cloud-related knowledge centered on AWS, Azure, GCP, etc.

◆ Treatise comprehension In
the AI ​​industry, where information disclosure is active, research skills and the ability to read documents (treats) are also required.

3.2 Key Career Paths to AI Engineers

◆ Learning specialized knowledge at a graduate school of information science If
you obtain a doctorate at a graduate school after graduating from university, you will be able to open the way as a front-line AI engineer.

◆ For those who have development experience as IT engineers studying at vocational schools and schools in the AI ​​field, vocational schools
where you can learn specialized knowledge such as machine learning and deep learning, which are essential for AI engineers, are now being seen.
In addition, an increasing number of schools are learning Python and R programming languages, which are indispensable for AI development.

◆ Learning through training within the company where
you got a job Many companies are focusing on AI development, and they are not catching up with the acquisition of AI human resources from the outside, but are strengthening the in-house training of AI human resources at a tremendous speed.
・ Implementation of AI technology courses and practical AI utilization exercises in collaboration with universities
・ Participation in external seminars and courses ・ Implementation
of technical exchanges and study sessions with external experts
・ Establishment of in-house study sessions and training

4. Summary