Institute for System Programming: How to Educate Research Engineers for IT

Viktor Ivannikov

Member of the Russian Academy of Sciences, Scientific Head of ISP RAS, Head of System Programming Chairs at NRU HSE, MIPT and MSU

 

Victor Kuliamin

Leading Researcher, Associate Professor of System Programming Chairs at NRU HSE and MSU

 

Alexander Petrenko

Head of Software Engineering Department, Professor of System Programming Chairs at NRU HSE and MSU

 

The existence and development of modern society significantly depend on technological progress. The latter is based on systematic production of innovations — technologies, which are socialized and embedded in a relevant niche for active use. Such innovation development is only possible when social institutes of science, industry and education function coherently. Relations between science and technology are often presented as a chain of knowledge transfer described by Francis Bacon (Fig. 1) ^[1].

 

A number of improvements of this scheme has been suggested, e.g., inclusion of feedback influences, acknowledgement that the development of each element is ordinarily determined by internal processes and only rarely involves the results or knowledge from others ^[2], and acknowledgement of heterogeneous character of the fundamental/applied science division ^[3]. Nevertheless, no serious researcher questions the necessity of coherent development of all the three its elements for stable long-term progress. Discussing a strategy of researchers’ education, the same long-term considerations require to take into account the relation between science and innovations on the one hand and education institutions, which ensure their reproduction, on the other.

Principles of Research Engineers Education

Research engineers’ education is one of the main challenges for creating a self-reproducing mechanism of innovation development. Research engineers are the personnel of the middle element of the aforementioned scheme, so they are mostly responsible for coherent development of all its parts. This challenge was recognized during the industrial revolution, and specialized technological universities were founded to cope with it. They were aimed at educating professionals who would be able to create new technologies.École Polytechnique (established in Paris in 1794) is an important example of such a university. Formally its priority can be challenged by the School of Mining established in 1735 in Selmecbánya, Hungary, which trained mining engineers (and which can be considered a predecessor of the University of Miskolc), andTechnische Universität Braunschweig founded in 1745. However,École Polytechnique was the first to use the essential principles of education of research engineers, “high-class engineers” who would be able to apply latest research in their real-life work and to conduct research themselves in the areas with lack of knowledge necessary for designing efficient solutions. These principles included

1. Integration of academic studies, deep understanding of fundamental disciplines and real practice;
2.  Intention to use gained knowledge to solve practical tasks that arise in the industry, military sphere or state administration;
3. Step-by-step systematic selection of the most talented and well-performing students (irrespective of their social status or wealth);
4. Elitism in education which allows to maintain high quality standards for decades.

These very principles were reproduced by Wilhelm von Humboldt in the course of higher education reform in Prussia and while establishing Berlin University in 1810, which was later given his name ^[4]. The first Russian engineering school was Moscow School of Mathematics and Navigation founded by Peter the Great in 1701. But some of the École Polytechnique principles were used only in the Institute of Railway Engineering Corps founded in 1810 by Agustín de Betancourt on the order of emperor Alexander I (now Saint Petersburg State University of Communication).

The “Phystech System”

In the 1950s, when Soviet economy needed quick restoration after war and new advanced and science-driven industries needed to be developed at the same time, the country’s leading scientists (P.L. Kapitsa, M.A. Lavrentiev, S.A. Khristianovich, S.A. Lebedev) initiated the foundation of Moscow Institute of Physics and Technology (MIPT, also informally known as Phystech) [5, 6]. One of its founding principles stated that in order to educate highly qualified professionals, students should participate in the work of research teams dealing with practical industry-related tasks. This became a basic principle of the so-called “Phystech System,” as part of which third year students (and later) communicate with scientists actively involved in strategic research projects and gradually are involved into such projects themselves.

As a result, MIPT graduates are ready for professional research, so for them post-graduate education is not just an introduction to research work but the most academically active period that ends with both a mature thesis and practical application of its results. The efficiency of such approach is proved not only by the achievements of the Soviet defense industry but also by the fact that the results of MIPT graduates are highly appreciated by the international academic community and the Nobel Committee in particular.

The world’s most famous technological universities (Stanford, MIT, Berkeley, Carnegie Mellon) cooperate with the industry too by collaborating with R&D departments of commercial titans (such as Microsoft, IBM or Intel) and public research centers such as INRIA institutes in France or Fraunhofer institutes in Germany. The complexity of integrating education and research in order to achieve practically important results, especially in the newest and fast-growing technological fields, makes it difficult to educate highly qualified professionals within traditional university framework.

ISP RAS: Integrating Research, Education and Technologies

The Institute for System Programming of the Russian Academy of Sciences (ISP RAS) is also using the “Phystech System” for more than 20 years to educate research engineers. ISP RAS was established in 1994 drawing on the experience of Lebedev Institute of Precise Mechanics and Computer Engineering of the Russian Academy of Sciences and the Institute of Cybernatics of the Russian Academy of Sciences, which were both exploiting the same model.

ISP RAS activity is best described as “industrial research,” i.e., work aimed at transferring research results into the industry (or other sphere of application). This means that the institute’s other activities are focused on ensuring that the technologies, software products and methods developed to solve various system programming problems meet all modern requirements and are ready for industrial applications.

Therefore, the main emphasis of ISP RAS activity is on the central element of the chain that connects fundamental science and application of new technologies. However, the other two elements can be found at ISP RAS too in order to ensure long-term operating capacity of innovation development mechanism.

Fundamental and experimental research are necessary to ensure that the institute’s evolution is in line with the most recent ideas and arising technologies. Ideas for new projects also originate from fundamental research. As technologies are developed to become true innovations and brought to active real-life applications they are also adapted for software products and services that can be used in practice too.

ISP RAS: Results Implementation Principle

ISP RAS does not launch start-ups because despite they can provide highest concentration on practically useful implementation of technologies, it could also destroy the institute’s research team and hinder the process of training new specialists.

Instead, the institute prefers to deploy its technologies and solutions through big industrial and applied research organizations, which at the same time exploit new technologies themselves and act as ISP RAS partners in promoting its achievements for broad use. Our foreign partners in the sphere of innovative activities are Samsung and Linux Foundation; Russian partners include State Research Institute of Aviation Systems and Kvant Research and Production Enterprise.

Besides research and technological work, we also take care of stable and smooth team development and renovation. Stability allows us to build long-term relationships with partners and customers, otherwise it would be difficult to reach and maintain a high technological level of R&D. Such stability is based on the research school that includes the institute’s leading researchers. It allows new employees to enter smoothly into the team and embrace both formalized knowledge and informal rules and behavioral patterns in order to work productively with other colleagues and to conduct high-level research on a systematic base.

ISP RAS: Reproducing Research Engineers

ISP RAS personnel renovation is ensured by participating in education. Each year 30-40 undergraduates from MIPT, MSU Faculty of Computational Mathematics and Cybernetics and — since recently — HSE Faculty of Computer Science join the departments organized with collaboration of ISP RAS. The departments were chaired by the institute’s founder and RAS member prof. V.P. Ivannikov until his death in November 2016.

The third year students begin attending lectures (3-4 per week) conducted by ISP RAS staff, participating in research workshops and getting acquainted with the work of various ISP RAS departments. One year later they start participating in specific research projects. Many students already have academic publications and become real specialists in their field by the second year of their master education. During their post-graduate studies they continue accumulating practical experience and deepen their understanding of the chosen specialization.

Besides that, post-graduate students also start teaching. They lead workshops and lab classes for students, conduct specialized courses, supervise students’ yearly projects and bachelor’s theses. Systematic accumulation of knowledge and experience over the years allows to nurture regularly high-class specialists who have valuable skills and have achieved substantial results.

Trying to separate practical specialists from teachers often results in the fact that the students who have “successfully” mastered the curriculum are not ready for real research work. They are good at solving well-formulated tasks but cannot cope with real-life difficulties in a situation when success depends on both being able to find the right solution and to formulate correctly the task itself. In the Russian context excessive preoccupation with academic mobility (in terms of domains and workplaces) often leads to greater time of incorporation in a new team and weaker research results.

Research projects that are directly relevant for industry needs allow ISP RAS to offer its employees competitive salaries comparable with high tech IT companies. Of course sometimes people change jobs or even move abroad but the institute manages to avoid massive “brain drain” due to high professional level of research.

So, people who are interested in cutting-edge research are just enthusiastic about working at ISP RAS. Having an interesting job means working on truly challenging problems, which implies dealing with innovative technologies and being able to look beyond the boundaries of existing knowledge. Moreover, true scientific research means openness of results and “visibility” of their authors, which often contradicts IT corporate policies. For ISP RAS openness of results and usage of open software both stimulate quality of work and help to promote new technologies that are being developed. Openness means that even young researchers working in big teams become known within the international IT community and build up reputation as world-class experts in their specialized fields.

Education of research engineers is one of the most important ISP RAS activities, all of which are closely integrated and provide a lot of feedbacks to each other; each component is important for the effectiveness of others. Here are some examples of such feedbacks:

• Research projects require effective professional education;
• Education of high-level researchers requires projects with important real-life tasks that allow their participants to master and develop cutting-edge technologies;
• Research that ends in the development of new technologies needs industrial projects to demonstrate efficiency of its results;
• Attracting new customers and partners and gaining new resources for development depends on research success.

ISP RAS organizational model may seem cost-inefficient. One could substantially cut the resources spend on teaching, participation in conferences, organization of various academic events or academic publications, concentrate on profitable projects and abandon unprofitable ones but that would cause institutional degradation. Current model enables smooth long-term development of the institute, it helps to build up capabilities to solve new problems of system programming domain and educate world-class experts.

Apparently, such model is not universally applicable, it would be difficult to reproduce it in many organizations. However, it facilitates successful education of research engineers who are able to develop new technologies systematically and to apply them in industrial practice.

Education of Research Engineers

To sum up, there are three key elements of the IT research engineers’ education strategy. First of all, such a strategy should be developed with long-term goals in mind. Education success criteria should be based not on formal process-related criteria but on an expert community evaluation of graduates as researchers. Actual education of a high-end specialist who has both broad knowledge in IT domain and deep understanding of his own special field requires 8 to 10 years.

Secondly, active participation in both cutting-edge research projects aimed at solving practical problems and teaching younger generations is a basis of research engineer education. This helps to develop an important complex of competences that cannot be measured by formalized criteria (such as citation index or other bibliometric indices). It is also important that by studying within a live research school with several generations of active researchers, students have a chance to witness a broad range of behavioral patterns at various stages of academic careers and adapt them to consciously build their own ones.

Finally, when students (yet lacking broad professional knowledge) shape up their individual curricula and often change the subjects they study (which of course helps to improve academic mobility indicators), it is impossible to neither train mature high-quality experts nor meet the country and economy long-term needs in various specialists. The strategic goal should probably be supporting teams that create an appropriate environment for education of such specialists, i.e., research schools and research centers that systematically conduct research and apply its results into practice while nurturing new academic, engineering and teaching staff.

References

1.      Bacon F. Of the Proficience and Advancement of Learning, Divine and Human. 1605.

2.      Kline S. J., Rosenberg N. An Overview of Innovation. In Landau R, Rosenberg N., eds. The Positive Sum Strategy: Harnessing Technology for Economic Growth. National Academy Press, Washington D.C., 1986, pp. 275-306.

3.      Stokes D. E. Pasteurs’ Quadrant: Basic Science and Technological Innovation. Brookings Institution Press, Washington D.C., 1997.

4.      Nipperdey T. Deutsche Geschichte 1800-1866: Bürgerwelt und starker Staat, vol. 1. C. H. Beck, 1983

5.      [In Russian] Khristanovich S.A., Lavrentyev M.A., Lebedev S.A. Pressing Challenges of Organizing Research Work. Pravda, 14 Feb 1956.