Konstantin Kanishchev, Developer in Grenoble, France
Konstantin is available for hire
Hire Konstantin

Konstantin Kanishchev

Verified Expert  in Engineering

Full-stack Developer

Location
Grenoble, France
Toptal Member Since
October 24, 2013

Konstantin is a full-stack software developer with a background in theoretical physics. He is an expert in modern C++, Python, and JavaScript. He is also an expert in functional the following programming languages: OCaml, Haskell, and ReasonML. He has deep experience in research-level software development, machine learning, big data analysis, and data visualization. Konstantin has high-level expertise in statistics, CS, and applied mathematics.

Portfolio

Univalence Consulting
Large Language Models (LLMs), Artificial Intelligence (AI)...
CERN (AMS collaboration)
OpenStack, Python, C++, Linux, Big Data Architecture, Data Science...
CERN (CMS collaboration)
OpenStack, Python, C++, Linux, Big Data Architecture, Data Science, Fortran...

Experience

Availability

Part-time

Preferred Environment

Git, Linux, C++17, Python

The most amazing...

...thing I've built is a LLVM Kaleidoscope language tutorial that's adapted for Boost Spirit, a C++ library used to develop parsers for text formats.

Work Experience

CEO

2019 - PRESENT
Univalence Consulting
  • Designed and deployed a back end for performant data retrieval and tokenization of NLP tasks. Further developed it into a full-stack web and mobile application.
  • Developed a specialized LLM sampling algorithm for specific data retrieval tasks together with a corresponding fine-tuned language model. Deployed it for large scale data processing (millions of records per day).
  • Managed a small R&D team in the area of LLM product research, focusing on performance optimization and reliability of the LLM responses.
Technologies: Large Language Models (LLMs), Artificial Intelligence (AI), Artificial Neural Networks (ANN), Full-stack, FastAPI, Node.js, Ruby on Rails 4, JavaScript, Linux, Python, Big Data Architecture, Deep Neural Networks, Natural Language Processing (NLP), GPT, React, Vue, Functional Programming

Research Fellow

2013 - 2019
CERN (AMS collaboration)
  • Participated in the development and maintenance of the data processing pipeline for the cosmic ray detector on the International Space Station.
  • Developed a MCMC statistical toolbox for analysis of science data.
  • Deployed a RabbitMQ cluster on OpenStack for monitoring the calibration status of a detector.
  • Had a Data Lead position in the Payload Operations and Control Center.
  • Worked on a worldwide international collaboration, joining CERN and NASA efforts on the most complex particle physics detector in space.
Technologies: OpenStack, Python, C++, Linux, Big Data Architecture, Data Science, Deep Neural Networks, Physics, React, Node.js, FastAPI, Full-stack, Artificial Neural Networks (ANN), Functional Programming

Graduate Student

2010 - 2013
CERN (CMS collaboration)
  • Developed and deployed algorithms for data analysis at the Compact Muon Solenoid experiment at the Large Hadron Collider.
  • Worked on software for heavy data analysis with Worldwide LHC Computing Grid. Developed a GRID job-management software dedicated for a particular CPU-critical task.
  • Collaborated on the application and the machine-learning algorithms for event selection and improvement of the sensitivity of the particle physics experiment.
  • Was heavily involved in statistical analysis of the retrieved experimental data.
  • Worked in a large (about 4,000 personnel) international and rapidly changing collaborative environment doing leading-edge fundamental research.
Technologies: OpenStack, Python, C++, Linux, Big Data Architecture, Data Science, Fortran, Physics, Full-stack, Artificial Neural Networks (ANN)

System Administrator

2005 - 2009
Institute of Mathematics SB RAS
  • Created and supported a small network of about 10 Windows desktops and a Linux server used as a computing workstation.
  • Created a simple Bash-script-based job management system.
  • Worked on cross-compilation and interoperability of legacy code, written in C, C++, Fortran, and OCaml.
Technologies: OCaml, Fortran, C++, Bash Script, Linux

Instructor

2004 - 2009
Novosibirsk State University
  • Taught courses in: C++ and MFC, Object Oriented Programming and Design, Numerical methods, and Symbolic Calculation with Wolfram Mathematica.
  • Active in more Physics related courses as well, such as classical and quantum mechanics, statistics, and cosmology.
Technologies: Object-oriented Programming (OOP), C++, Physics, Functional Programming

Intern

2004 - 2005
Novosoft
  • Worked in a small team of software developers.
  • Designed and worked on the implementation of software for testing of video codecs.
  • Implemented a library for reading various image and video file formats. Due to the absence of documentation, this sometimes required reverse-engineering of the files.
Technologies: Microsoft Foundation Classes (MFC), Microsoft Foundation Class (MFC) Library, WinAPI, COM, C++, Functional Programming

2HDM explorer

Getting an intuition about multivariate data is a common problem in vast majority of applications. This project is a presentation of results of complex calculations for a particle physics experiment. Given a set of parameters of a certain model one wants to see what kind of phenomena are expected and which regions of parameters are responsible for which outcomes.

MPM Flexible Body Simulation

http://physics.stackexchange.com/questions/629/why-does-one-experience-a-short-pull-in-the-wrong-direction-when-a-vehicle-stops/6107#6107
Using only academic papers, I've implemented from scratch the Material Point Method simulation of a deformable body in C++. Includes QT + OpenGL GUI.

LLVM Kaleidoscope tutorial in Boost::Spirit

https://github.com/KKostya/SpiritKaleidoscope
The LLVM compiler tutorial uses a simple hand-written lexer and parser, implemented either in C++ or OCaml. The project is an attempt to use Boost::Spirit library as a parser for that tutorial.

Realtime Joukowski map in JavaScript.

http://bl.ocks.org/KKostya/6066548
Conformal mappings is a powerful mathematical physics technique for solving 2D boundary problems. I always loved the way the result look. So I've implemented it using JavaScript with the help of the d3.js library.

Languages

JavaScript, OCaml, C++, Fortran, Python, Bash Script, XML, C++17

Frameworks

Boost, Flask, Flex, Ruby on Rails 4

Libraries/APIs

Standard Template Library (STL), D3.js, Node.js, Vue, React, WinAPI, Microsoft Foundation Class (MFC) Library, jQuery, Microsoft Foundation Classes (MFC)

Paradigms

Object-oriented Programming (OOP), Functional Programming, Template Metaprogramming, Data Science

Other

Software Development, Physics Simulations, Big Data Architecture, Physics, Artificial Neural Networks (ANN), Full-stack, FastAPI, Natural Language Processing (NLP), GPT, COM, Deep Neural Networks, Generative Pre-trained Transformers (GPT), Large Language Models (LLMs), Artificial Intelligence (AI)

Platforms

Linux, OpenStack

Tools

Git

2010 - 2013

Ph.D. in Theoretical Particle Physics

University of Trento - Trento, Italy

2005 - 2008

Master's Degree in Physics

Novosibirsk State University - Novosibirsk, Russia

2001 - 2005

Bachelor's Degree in Computer Science

Novosibirsk State University - Novosibirsk, Russia

NOVEMBER 2020 - PRESENT

Natural Language Processing Specialization

Coursera

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring