# Zhuykov Alexandr Victorovich

WhatsApp [phone], Telegram @icivilab или [phone] или https://t.me/icivilab
tel: [phone].

https://vk.com/icivilab
https://www.linkedin.com/in/tatoy/

Saint Petersburg, willing to relocate (Moscow, Belgium, Great Britain, Spain, USA, France, Austria, Germany, Switzerland)

TensorFlow, Keras. machine learning researcher + huge, very different IT experience
Management
IT, Internet, Multimedia
Manufacturing, Technology
Science, Education
Employment: work placement, volunteering, part time, project work, full time

Work experience 30 years 10 months
August 2015 — currently
6 years 3 months
Neural networks for science and business
Saint Petersburg

researcher. neural network models and teaching methods.

creating a kernel that can be used in
• oil and gas production
• banking
• in search engines
• in medicine, treatment of cancer, viruses, etc.
• air and railway transportation
• games
• bots
• further development of universal AI
• and others. The scope is very wide

more specifically:

• neural network models and teaching methods.

Models are universal, in continuation:
• Advanced, deep understanding of video, photo, sound and other streams.
• Solution of complex problems by a neural network, non-standard issue of solutions and solutions requiring long sequences of solutions (automatic, composite).

At this stage, I use for this
• TensorFlow, Keras, Python.
• TensorFlow Hub, Universal Sentence Encoder, Token based text embedding,
• Trained and ready-made neural networks,
• Datasets,
• Teaching a neural network without a teacher 2 on unlabeled texts and documentation,
• Automation and robots in Python.
• and many others. other.

Models are partially developed, passed some performance checks, busy with implementation, code in colab.research.google.com and a couple more places with Tesla + in Python editors on PC.
Final part, experiments, performance check, testing, preparation for demonstrations.

Creativity. Special systems for generating ideas and solving problems.

New product for business.
Significant product improvement.
I will create a great competitive advantage.
Neural networks for science.
Researcher Artificial Intelligence.
I make custom neural networks that are up to 10 times smaller and faster.
I raise the Test set accuracy to 99%.
I develop a versatile and strong AI.
I design models of artificial intelligence, network architecture.
TensorFlow, Keras and any others.

I develop advanced algorithms for neural networks, create automatic neural networks.
Systems on non-networks.
Software robots. automation.
January 1991 — currently
30 years 10 months
Brain Research. On my own, independently
Researcher on memory, the human brain. Brain Research
Futurologist.
Systems for Memory. Decision and idea generation machines.
Using features in human evolution to generate ideas and solutions.
Researcher. Research, findings, analysis on memory, the brain and human ability development.
March 2000 — september 2015
15 years 7 months
On my own, independently
Moscow

Programmer, lead developer, IT support
Programmer: robots, kernels and automated systems, web projects, websites, IT work.
Code, Video, Graphics and sound, hardware, operating systems, kernels
Key skills
Python
TensorFlow
Keras
Big Data
Data Scientist
Machine Learning
нейросети
Deep Learning
Scikit-learn
AI
Artificial intelligence
Artificial neural network
Neural network
Git
GPT-3
Transformer
BERT
Attention
ELMo
RNN, GRU, LSTM, Autoencoders (VAE), Generation - GA
Convolutional NN, Autoencoders, GAN, Seq2seq
Jupyter Notebook
K-means
k-nearest neighbors algorithm, k-NN
Artificial Intelligence Institute
Neural networks
Universal Artificial Intelligence
Universal AI

it's an automatic translation. my native language russian:
https://spb.hh.ru/resume/c9a2ab8fff094e27fe0039ed1f636c62427275

core, a universal, automatic system that is rapidly gaining momentum and creates a great competitive advantage, programmers and neurons, neural network modelers.

Therefore, I have provided a demonstration of a neural network a year ago below and I can provide a beautiful graph of the profitability of trading on the exchange.

I am creating a kernel that can be used in
• oil and gas production
• banking
• in exchange trading
• in search engines
• air and railway transportation
• in medicine, treatment of cancer, viruses, etc.
• games
• bots
• further development of universal AI
• and others. The scope is very wide

more specifically:

• Developing neural network models and teaching methods. The goal is an expanded, deep understanding of the text, incoming information and relevant, accurate results.

Models are universal, in continuation:
• Advanced, deep understanding of video, photo, sound and other streams.
• Solution of complex problems by a neural network, non-standard issue of solutions and solutions requiring long sequences of solutions (automatic, composite).

At this stage, I use for this
• TensorFlow, Keras, Python.
• TensorFlow Hub, Universal Sentence Encoder, Token based text embedding,
• Trained and ready-made neural networks,
• Datasets,
• Teaching a neural network without a teacher 2 on unlabeled texts and documentation,
• Automation and robots in Python.
• and many others. other.

Models are partially developed, passed some performance checks, busy with implementation, code in colab.research.google.com and a couple more places with Tesla + in Python editors, php on PC. If possible, I delegate the work to programmers.
Final part, experiments, performance check, testing, preparation for demonstrations.
'' '
• Creation of new neural networks and non-standard models, closer to AI.
• Systems on neural networks for science, product and business, from start to finish.
• I make your neural network up to 10 times faster and smaller.
• Raise Test Accuracy to 99.5% of your neural network (where often a neural network is more accurate than a human).

For an example from my Kwork:
... and so on up to 99.5%.

Below is my demonstration - acceleration of a neural network by 70 - 80% in work and repeatedly in training. (+ with improved performance and reduced up to 10 times)

tel: [phone],
WhatsApp [phone]
Telegram @icivilab or Telegram [phone]
vk - icivilab
Zhuikov Alexander - Tatoy (icivilab - @ Syrius4 - Τατόι)

Demonstration:
Recognition of voice commands.

Outcome:
My Model1 is 11x smaller than the manual (count_params 148.645 vs 1.625.611)
And much better in all respects.

more convenient + images here
https://hackmd.io/@syrius4/neural

Photos and descriptions to them in detail right there. (see portfolio, images to resume)

Indicators of this guide as of 03/21/2016:
https://www.tensorflow.org/tutorials/audio/simple_audio
Total params: 1,625,611
val_loss: 0.4848 - val_accuracy: 0.8388
Test set accuracy: 84%

my Model1
count_params 148,645
evaluate loss 0.3729 - evaluate accuracy 0.88
Test set accuracy 84.62%

my Model2
count_params: 3,003,038
evaluate loss: 0.2823 - evaluate accuracy: 0.9087
Test set accuracy: 89.00%

The same 6400 training sound files / 6400/800

This algorithm was invented a year ago. He's different now.
The algorithm is universal, for any neuron. This is just an example of simple audio, quick.

My algorithms are not only for collecting neural networks and optimizations.
They are more for creating systems of a new generation, extended, universal, etc.

I design and develop new neural networks.
Robots, systems, speech, text, video.

Understanding technology or practiced, among others:
• TensorFlow and Keras (functional API and Sequential)
• Preparation of datasets, data normalization
• Custom Embedding, including invented copyright technologies
• Freeze weights and layers.
• Saving - restoring the model and weights.
• Transfer Learning.
• Pretrained networks.
• Custom layers,
models, custom activation functions and metrics.
• Avtotyunig, AutoML.
• Regularization and Dropout - fighting overfitting.
• Batch Normalization.
• Masking.
• Manipulations with complex LRs.
• use in TF of 16-bit and 32-bit types (float16, float32) for acceleration or precision.
• TensorBoard.
and other.
• Classification, regression, generation, etc.

Understanding technology
• RNN, GRU, LSTM,
• Convolutional NN,
• Encoding-Decoding, Variational Autoencoders (VAE),
• Generation - GAN,
• Seq2seq and Attention,
• Transformer BERT, GPT-3,
• ELMo and more.

Clustering:
• K-means
• Affinity Propagation
• Mean Shift
• Spectral clustering
• Hierarchical clustering
• DBSCAN
• OPTICS and more.

Universal, Strong AI (both from real to real and Futurology (future, ideas without criticism).
Unsupervised learning 2.
Part-teacher training, Semi-supervised learning.
Technologies and inventions of data supply for neurons, including "on the fly" (video, text, sound, speech, documentation).
Training on unlabeled data.
AI self-creation, "Snowball", self-improvement, self-learning.
Automatism and self-optimization.
An expanded picture of the object, a complex neural network.
Complex sequence, pipeline, chain for a neural network.
Exponent in AI.
He modernized and invented regression and prediction systems (in the past - a robot on the Moscow and NY stock exchanges, cryptocurrency.

Development of the system of the Institute of AI itself and searches, works, the latest production (intelligence).
Development of research systems (and a special Institute)
Speed ​​education developer.

as well as
Hundreds of discoveries, ideas, own custom technologies and directions of research, researches.
I am a generator of ideas. For this I have my own big system and a lot of practice. Yes, my brain is wired with algorithms for generating ideas in practice.

Favorite direction - development, new research and development, large-scale progress, civilization, the latest industries, the development of medicine and science on a large scale.
Systems for AI, thinking, solutions, inventions, production of the latest technologies and products.
Health and life extension, support of the mind, memory and all processes in the body in excellent condition, breakthrough technologies and products.
• Google Colaboratory, Jupyter Notebook and PC code editors, python.

• Large independent education in the direction of 'Startup'. One of the most progressive in terms of generating ideas and solving problems.

until 2010
• Successful market maker (in% - always profit, robots and neural networks for risk-free trading).

My systems trade and never stop for many years. I do hard core experiments all the time and always do everything. If the robot's money management allows you to drain everything, then it will.
Experience in forecasting rates of currencies, stocks, cryptocurrencies. Correlations.
• Created robots-market makers at the Moscow Exchange (Quik) and at crypto-exchanges - more than 3 years of experience. This is the last work, the result is profit, smart, they count 1,000,000 combinations in minutes on i5, 3 levels of cache, protection against hairpins, from external API glitches.

Computer architecture, computer hardware. My first computers long before Pentium1 on green screens and the advanced practice of assemblies and upgrades on motherboards, caches, chips, screws, memory and peripherals.

Creation / Programming:
• sites, shops, robots, services, Internet project management.
• Python, php, lua, qlua, mySQL, PhpMyAdmin, json, xml, api client, Vk api - (more than 3 years experience), did everything himself, and also attracted masters, organizer, idea, project leader, own projects.
• php OOP, python OOP, WebSocket, JS - just a little bit about that.
• I will master any language and framework - quickly.
• charts, informers and calculators - cache, protection from hairpins, from overload, from robots, keeping access to hidden search engine robots, handling loss of connection with an external API, using a php library, using js and whatever.
• Participated in project management (creation of robots, websites for the company, design).
• Bootstrap3, Bitrix, 1C - it was a long time, experience less than a year.
• Databases are not a bad idea. If not large, I quickly collect my bases in php and python for the cores of robots and neurons -
much does not go beyond the mem and L1,2,3 cache of the processor or GPU, smartly - without SD, HDD or network.
• Experience with a client (trust, personal location).

• statistics, Google Analytics, Yandex.Metrica, LiveInternet, Wordstat.
• Search engine optimization sites, the site has been profitable for 10 years without support. The average was 2000 unique per day, max 20,000.
• Yandex.Direct, Google AdWords and Target.
• Yandex Advertising Network, Adsense.
• CRM (wrote CRM to myself in php + Vkontakt API).
• Advertising banners.

Internet shops
• experience in an elite boutique: new website, photo, video, net, all IT, image maker, designer, tailor, fashion.
• OpenCart. Auto-filling, loading-unloading-parsing of goods from other sites, synchronization with 1C
• CMS.

Radio engineering, radio electronics, equipment - soldered, assembled according to schemes, repaired until 1992
Car enthusiast, assembled - disassembled a car.

Photo and video filming, graphics, video:
• Portrait, backlit, no shadows, movable platform and tripod.
• CorelDRAW, Photoshop.
• 3D editors.
• Video editors.
• Design and Web Design.

Sound, voice acting and music.
• Cubase. Fruits. Cakewalk Sonar, Waves, Sound Forge, MIDI and VST plugins - Great experience.
• effect processors, sound processing, samples and synthesizers, mixing.

Systems, OS, Office
• installed / configured Windows, office networks and the Internet. Great experience.
• Computer hardware, repair, office network setup.
• Linux, Ubuntu, Debian, Apache / Nginx / php / http-server.
• Works on FTP.

• English - 10 years study of scientific documentation, technical and miscellaneous, with a wide coverage. And yet, neural network modeling kills English and everything else with a margin)

Positive, enthusiastic.
WhatsApp [phone]
Telegram @icivilab or [phone]

Russian — Native