Salaries in AI: where more money and who are looking for in Russia

    Specialists in artificial intelligence are paid almost twice as much as other IT professionals. We figured out what salary you can expect in different areas of AI in Russia, who are looking for Yandex, ABBYY and Sberbank, and what courses can be used for training in this area.

    How much does an AI specialist earn in Russia

    The salary of a specialist in artificial intelligence in Russia, according to HeadHunter at the beginning of 2018, was about 190 thousand rubles. This is twice as much as the average salary in IT, which at that time amounted to approximately 90 thousand rubles. By the end of the third quarter, it almost did not change .

    The most promising area in 2018 remains Big Data: experts in this field are offered a salary of around 200 thousand rubles. There are about 180 thousand machine learning specialists, more than 140 thousand in the field of neural networks.

    The number of vacancies in each of these areas is growing faster than the number of resumes - this is typical of the situation with personnel in IT as a whole. According to the researchIIDI, by 2027 in Russia there will be a shortage of about 2 million IT specialists.

    According to the IIDF forecast, by that time artificial intelligence, machine learning, big data analytics, computer vision and augmented reality will be among the most sought-after areas.

    Who are looking for in large companies

    Artificial intelligence is used primarily by large companies such as Yandex, Mail.Ru Group, MegaFon, MTS, Beeline, Tele2, ABBYY and Sberbank. What is it for some of them and whom they are looking for:

    1. ABBYY

    ABBYY is one of the world leaders in intellectual data processing and linguistics. Its AI-based solutions allow you to recognize text data, work with printed documents and PDF files, conduct semantic search, and also find translations of unfamiliar words and phrases.

    One of the main achievements of the company is the Compreno system, which allows you to analyze and understand natural language text. ABBYY specialists worked on the creation of this system for about ten years, the cost of the project exceeded $ 80 million.

    Compreno can be used, for example, to systematize archive documents: with its help, it will be possible to find information by fields or details, as well as by text.

    Who is looking for ABBYY: now the company is neededData Scientist - for experiments and prototyping in the field of word processing (NLP) in the advanced development department. The candidate is required to have knowledge of machine learning methods and neural networks, algorithms and data structures, Python programming experience and some other parameters.

    2. “Yandex”

    The largest Russian search engine has been using AI technologies for several years in its search engines. For example, in Yandex.Dzene, this allows for the issuance of personalized recommendations of content in accordance with the interests of the user.

    “In many ways, it is similar to a search engine. Only if the search is looking for something specific, does Zen respond to a broader query: what is interesting to a specific person, ” said Viktor Lamburt, head of Yandex.Dzen, at the start of the service.

    Who is looking for "Yandex":Right now, the company needs a machine learning developer for Zen, who will collect data, train models, evaluate them in experiments, and write production code. “First of all, we expect good knowledge of machine learning and statistics from candidates, but industrial development experience will also be a big plus,” the vacancy says .

    3. "Sberbank"

    The direction of machine learning and artificial intelligence has been developing in Sberbank since 2013. The main goal is the creation of new intellectual products and services for both internal and external clients, as well as the optimization of banking processes using machine learning technologies.

    So, in early 2018, the bank launched the first neural network in Russia to evaluate commercial real estate. AI allows the bank to almost instantly assess the collateral. This AI works with a regularly updated street retail database. This database is replenished from several types of sources and contains the main characteristics of analogous objects, their photos and prices.

    The neural network receives the characteristics of the object, which must be compared with others, and on the basis of the collected data selects the closest analogues that are used to calculate the cost. If experts need hours and even days for this, then neural networks take several seconds to analyze.

    Who is looking for "Sberbank": now the bank needs several data scientists for different projects. For example, in one of the vacancies in Moscow, a specialist is required to have experience in solving
    data science problems for business, big data experience, good programming skills (Python, Spark, SQL) and knowledge of machine learning libraries.

    In addition to Sberbank, data scientists and machine learning specialists are required many other banks, including VTB, UralSib and BinBank.

    Where to start

    Only 30% of specialists working in the field of AI, studied machine learning or big data at the university. This is evidenced by the results of a survey of 16 thousand users of Kaggle, conducted at the end of last year. More than half (66%) of all respondents consider themselves to be self-taught: they used various courses to study new disciplines.

    Microsoft evangelist and head of AI School in the Binary District Dmitry Soshnikov outlines four main types of courses on the Russian educational market:

    • short courses on the role of AI in business - for managers who need to get a first acquaintance with the subject;
    • highly specialized courses such as "Recognition of images in five hours" - for those who want to work out specific skills;
    • classical universities - for those who want to get a detailed understanding of all algorithms and learn how to program neural networks independently;
    • long special courses for data scientists - for those who want to get a new specialization and completely change jobs after training (such courses last at least several months).

    Each type of course has its drawbacks. Executive courses, for example, are good for briefly acquainting with best practices in the field of artificial intelligence, but do not provide a complete picture and a general understanding of all the capabilities of AI and its limitations.

    The same problem with highly specialized courses: they do not allow the listener to form an understanding of the fundamental principles of the work of AI. The student can master some practical skills, but technologies become obsolete every six months, and skills - with them.

    Classical university courses for beginners may be too complicated: here we have to remember forgotten sections of mathematics. Future data scientists, in addition, usually need to have good programming skills.

    For developers who want to figure out how and why you can use AI in their company, the best course is not too long, but an intensive course that will allow you to learn how to solve typical problems. In AI School , for example, students learn five blocks of tasks per month:

    • typical tasks solved by pre-trained cognitive services (recognition of faces, emotions, voices, etc.). One of the homework - to make an application that recognizes the emotions of the main characters of the film in the course of the action;
    • creation of the simplest conversational AI;
    • classical machine learning tasks (demand prediction, predictive analytics, etc.);
    • work with images (classification, detection of objects) and video;
    • work with text and natural language (classification, generation, etc.).

    After that, the listener can already decide whether he needs to acquire additional knowledge in order to learn how to solve more unusual problems.

    According to Soshnikov, passing the course will not lead to an instant salary increase, but this will make the specialist more attractive to the labor market. And this will allow both to demand a raise from the current employer, as well as to look for other opportunities. It all depends on the person himself.

    Also popular now: