Download my resume in .pdf
Download my PhD and Education credentials in .pdf

November 2022 - present

AI Dialogue Facilitator / Data scientist / Data Analyst; independent contractor.

As an AI Dialogue Facilitator I engage with AI on behalf of a client. I select the most suitable models, methods of interaction, and implement necessary software tools. Subsequently, together with the customer I lead the dialogue, interpret and communicate interim outcomes, and iterate the process as needed to attain a concrete outcome.

AI Dialog Facilitator performs the following tasks:

  • Formulates tasks for the AI model(s) used by the company, ensuring a reasonable expectation of success.
  • Collects and organizes materials and data to provide descriptive context and documented precedents for interactions with the company’s AI.
  • Recommends the appropriate form of interaction with AI for each task, such as one-on-one dialogue, group conversation with AI, or multilateral conversation involving multiple copies of AI with adversarial instructions and one or more humans.
  • Sets up and moderates the course of all these types of interactions, documenting the stages of conversations, AI responses, and momentary decisions made by the human participants.


Game Insight, UAB

April 2020 - November 2022

Chief Data scientist / Data Analyst; independent contractor.

Detailed achievements in this position:

  • Developed a mathematical foundation for the analysis of Correlated Sequences, implemented the code and models, tested it on standard data sets and achieved state of the art or higher results (trade secrets of the company);
  • Launched on the Company infrastructure the Ax.dev Experimentation Platform for Bayesian Optimization, (developed by Meta platforms (former facebook Inc.) and conducted numerous online, online-offline and multi-objective experiments with modications of Company’s games. On this setup conducted ‘interventions’ (collection of counterfactual data) in order to nd causal relationships and achieve explainability. The micro-service is built on FastAPI, APScheduler, asyncio and uses Airflow as a secondary sub-service;
  • Set up and used the AWS infrastructure of the Company (SageMaker Studio, SageMaker, Athena, Glue, PySpark apllications) for Machine Learning, now I’m migrating to Databricks;
  • Implemented (Athena, SQL, Dask) multiple data extraction pipelines from a petabyte size Data Warehouse of the company to the data sets that were used by me for machine learning;
  • Acquired a hands-on experience of working with, AWS RedShift cluster because the main DWH of the company is built on it. The company maintained around a dozen mobile games with several million of active users each.


January 2017 - 2020

Data scientist / Data Analyst / Data engineer/ Data programmer; independent contractor.

I collect the raw data from your automated business-process or experimental set-up and transform it into a meaningfull model that lets you make decisions. The meaningful visualisation of data is a part of my job. Typically I take small to medium projects (less than a year) in Chicago, IL or tele.
Green Card holder, full work authorization in the US.

Tools: I feel most comfortable using Python / RStudio / R and SQLite (with an external 1TB HD on USB 3.0) or Apache Spark (on databricks.com), and remote databases for my everyday work but I can dive into the depth of LLVM and C++ too if it is (absolutely) necessary. Also, there is nothing in the world a Ph.D. in Physics and Mathematics who started writing code in assembler on PDP-8 in 1978 and has been working with all sorts and kinds of experimental data (from medical and physics experiments to business and market/financial data) ever since can not learn in a month. I can prove it if necessary.

A typical project:

  • Definition of the problem and tools; definition of what should be delivered;
  • Data cleaning and organizing in a usable ‘library’ (I don’t use marketing terms like ‘lakes’ etc.etc.);
  • Modelling and the ‘Design of Experiment’ or choice of a Machine Learning tool(s);
  • Work itself;
  • Presentation of results and deliverables; reprocessing and reformatting of results if necessary.

Interests: I’m interested in Topological Data Analysis (TDA), Multivariate Scaling and Principle Component Analysis. I read and write texts and code about it in my spare time. The ‘inverse problem’ of (re)constructing decision forests from a trained NN is very interesting too, but it is still a bit of a desire rather than a real work that can be done in a straighforward way.


On-line Card Technologies, Inc. (Chicago, IL)

2003-2016

President and CEO; sole owner.

Managed the design and development team of ON-LINE CARD ™ - standalone digital ID / payments processing / billing web-service.


ROBOXchange.com

2002-2004

Owner, CEO.

Fully automated e-currency exchange service. Led the design and development, built a business around it and managed the continuous support and development of the service, including the financial management and partnerships with all e-money services that existed at the time.


Guarantee Agency, JSC.

1999-2003

Co-owner, CEO.

Led the product design team and product implementation team of the first direct-access online exchange INDX.ru (still operates). The exchange traded derivative direct-lending instruments for e-money (webmoney.ru); devised the contracts that were traded on the exchange and added several important features to the e-money system itself. Tools: SQL.


A.V. Fedotov and Co., Inc.

1995-1999

President and CEO,

Licensed commodities futures brokerage firm, accredited with MICEX (Moscow). Devised and implemented a futures trading system based on composite indicators and Bayesian neural network. Tools: METASTOCK, Neurostock, Excel script, Visual Basic.


LYNX, Inc.

1992-1995

CEO,

Research and development company that engineered and patented an Erbium medical laser for blood sampling (scarificator).


Lebedev Physical Institute of the Academy of Science (Moscow)

1979-1992

Junior Research Associate - Research Associate;

Physical experiments and experiment automation on high-power relativistic electron beam accelerator. Skills: PDP 8 and 11, Apple ][, IBM PC, DEC VAX, CAMAC (Computer-Aided Measurement And Control), Tools: Assembler, Focal/Fortran, Pascal/Modula, C/C++.


Moscow State University (Moscow)

1973-1979

Student (Physical electronics, Plasma physics) - Ph.D. in Physics and Mathematics. Plasma physics.

My resume on LinkedIn