Hello there! My name is Yury Shimansky. I like cybernetics, biology, and mathematics, and I try to combine them in my theoretical research and practical applications. My double education, one in cybernetics and computer engineering and another, in neuroscience allows me to find novel, nature-inspired solutions to problems arising during development of automatic systems for information processing and optimization of control decisions. For example, I solved the long-standing problem of finding a biologically plausible procedure for learning in artificial neural networks.
I’ve published several papers on machine learning and artificial intelligence. Recently, I published a theoretical study in which I described a trans-algorithmically constructive model of biological systems. That model can be viewed as a new, exciting direction in AI technology, which can be especially useful in space explorations, where possibilities for manual control are limited and extra self-adaptive power is required.
I’ve successfully applied my skills in AI to the development of functional components of stock trading models and optimization of their parameters.
I was interested in cybernetics and biology already in high school. At that time, I lived in Kiev, Ukraine (I was born there, now I'm a US Citizen). After graduation, I decided to study cybernetics and computer science in Kiev Polytechnic Institute. There I became fascinated with artificial intelligence, and my computer program for pattern recognition won the first place in the competition between student projects. After graduation from the institute with “a red diploma” (equivalent of cum laude), I worked there on a challenging and exciting project: development of software for training pilots of space shuttles. This was rather interesting, but I continued going deeper into adaptive systems and artificial intelligence. I also began reading about the brain, the most complex naturally intelligent system, and rather quickly realized that I wanted to study it. Soon I had a lucky opportunity: I learned that in Bogomoletz Institute of Physiology there was a Ph.D. student position for someone who would be very good with computers and also interested in neuroscience. That was a description of me!
At the time of my starting a PhD program in Bogomoletz Institute of Physiology, there was only a computer center for offline data processing and simulation, but no labs had equipment for computer-based recording of experimental data or active control of ongoing experiments. There was a computer of PDP-11 type with CAMAC modules (ADC, DAC converters, etc.), and my first task was to interface that with the experimental equipment and write software for data recording, processing, and online control of experiments. I worked long hours every day and finished the task in a month or so. Then I had to learn how to perform electrophysiological experiments, expand the software for data analysis, write papers, and so on. Those were happy and enjoyable days of scientific research. After successfully finishing the PhD program and obtaining a graduation diploma, I continued my research.
Bogomoletz Institute of Physiology was well known internationally, and we had visitors from other countries. One of them, Dr. James Bloedel, an American neuroscientist from the Barrow Neurological Institute (Phoenix, Arizona) visited my lab and was positively impressed with the quality of the experimental equipment and research. He invited me to visit his lab. At the time of trip to the United States at the beginning of 1992, I thought I’d spend a few months there and then return to continue my research in Ukraine, which was progressing really well. However, the downfall of the Soviet Union and the steep decline of the economy in the former Soviet republics spelled doom for scientific research in Bogomoletz Institute of Physiology, and, therefore, I accepted Dr. Bloedel’s offer of a postdoctoral position at the Department of Neurophysiology and moved to the U.S.A. Eight years of very interesting and fruitful research followed.
In 2000, Dr. Bloedel decided to get a new position and move his labs to a different state. After much contemplation, I respectfully declined the opportunity to move there with him and instead took a Research Professor position at the Arizona State University (ASU). I felt that I had a lot of information about the functioning of the central nervous system obtained from numerous experiments, and it was time for me to start going back toward theoretical research. To begin, I wanted to concentrate on theoretical investigation of learning systems, and I used artificial neural networks to model learning to optimally control a limb-type object. However, despite my success in the theoretical area, getting sufficient financial support for theoretical research proved really challenging. I wrote several grants, but none of them were funded.
By that point, my friend and I already had a company called AlgoBrain for automatic stock trading. I’ve tried a few approaches to price move prediction, including usage of artificial neural networks. The results were encouraging, but that project was taking increasingly more time, distracting me from scientific research. Consequently, I put algorithmic trading on hold for a while.
One day in 2007, I was approached by a friend of a friend, who asked me whether I wanted to switch the main job and the hobby by getting a position at a just created, well-funded startup company for algorithmic high-frequency trading. Scientific research was supposed to become my hobby. It wasn’t an easy decision for me, but, after a period of hesitation and contemplation, I accepted. The company (SXP Analytics) rather quickly became profitable. My input turned out very useful, and during the next few years business was really good. Unfortunately, for reasons not related to the quality of the trading model, the company had to close in 2012.
Fortunately for me, the head of the department of strategic investment at Jefferies, LLC was looking to add high-frequency trading (HFT) technology to their long-term trading. Through a friend, he offered me a position of vice president, which I accepted. I developed software for HFT. However, around that time, as experts in the field know well, making good profits using traditional models for price move prediction became increasingly difficult. Hence, I had to make significant improvements in the trading model. Another problem was technical: HFT needs a very good infrastructure with minimal delays in reaction to public orders and sending out trade orders. Eventually, despite the fact that I had developed a profitable trading strategy, the department’s head decided that expenses were too high to continue the HFT direction, and it was closed at the end of 2016, resulting in the removal of my position.
I found another job in stock trading only next year, with Migdal Research. In a few months after that, it merged into Fresnel Research, where I got a senior quantitative data analyst) position. Although I developed a successful trading model, the chief software architect was more interested in pursuing his own ideas, and my time was more than fully occupied with helping him to program and debug his solutions. In the middle of 2019, due a decline in the chief software architect's health, he resigned, which resulted in the investors’ pulling their money out and disintegration of the company.
During recent years, I continued (as a very serious hobby) my research in the areas of artificial intelligence and neuroscience. I’ve written and recently submitted a paper on trans-algorithmicity of life emergence and evolution, and I continue working on my AI applications for stock trading. I have innovative ideas for practical utilization of trans-algorithmic AI technology, and I’m looking for investors and partners to put those ideas to use.