About | Ivan Yakovkin

Formal Education and Achievements

Faculty of Physics, Taras Shevchenko National University of Kyiv, specialization "Quantum Computers, Computing and Information":

2019 — Bachelor's degree with honors
2021 — Master's degree with honors
2023 — taught an advanced course on numeric simulations in COMSOL
Currently pursuing a PhD in the same field

Certificates and Awards:

2010 — graduated Kyiv Music School No. 8 with honors
2017 — Japanese-Language Proficiency Test (N5)
2019 — English for professional activities (Natural Sciences) certificate (C1)
2021 — FreeCodeCamp certificate on "JavaScript Algorithms and Data Structures"
2022 — Best poster award in the 7th International Symposium on Dielectric Materials and Applications ISyDMA'7
2023 — Arena FIDE Master title in chess
2023 — Scientific Excellence Award from the Taras Shevchenko National University of Kyiv

University

Research Interests

Sun

Nematic Liquid Crystals (NLC):

Analytical and numeric modeling of NLC director dynamics
Optics, photonics, plasmonics in NLC
Bi- and multistability in NLC; hysteresis of NLC director orientation
NLC-based tunable microantennas and lasers
Finite anchoring energy, pretilt, easy axis gliding, flexoelectric polarization

Solar Physics:

Magnetic fields and Doppler velocities in solar flares, prominences, and sunspots
Modeling and inversion of Stokes profiles under non-LTE conditions

Scientific Computing:

Advanced numeric simulations using COMSOL
Development of custom software using Python, Maple, Mathematica, Delphi, Matlab

Software Development

Experience in Finance and Credit Risk:

Forward and swap rate prediction using Variational Autoencoders (python: pytorch, numpy, scipy, pandas, plotly)
Implemenation and support of stress testing models in Wholesale Credit Risk (python: numpy, pandas, pandera, numba, scipy, numexpr)

Stocks

Created Products

Virtual Photometer

Virtual Photometer – a software for precise and semi-automated digitalization of spectral material captured on photo-plates. Virtual Photometer accounts for characteristic curves of both the scanner and the photoemulsion, performs data averaging, noise retraction, precise reference point selection, and many other features. Virtual Photometer provides a powerful graphical user interface, is written in Delphi, and is fully integrated into the data processing routine.

Profile Manipulator

Profile Manipulator – a solution for expedated processing, analysis, and visualization of spectral line profiles. Profile Manipulator allows to interactively visualize and process large amounts of spectral data, combine and calculate derivative profiles while keeping track of the measurement errors. Profile Manipulator is written in Delphi and is now an important part of the spectral processing workflow.

EYLEED – a software for Low Energy Electrod Diffraction pattern calculation and inversion within the kinematical approximation. EYLEED is written in C++ (originally in fortran90), and offers a rich graphical user interface, significantly facilitating the interpretation of the LEED patterns.

StatPrism – a software for general statistical analysis and automatic verbal desctiprion of the obtained results.

Solar Radiative Transfer (SolRaT) – a modular cross-platform code for non-LTE synthesis and inversion of Stokes profiles for arbitrary magnetic fields (python, status: approaching MVP)

Reflective-Transmissive Deep Neural Network (RTDNN) – a Machine-Learning-based tool for semi-automated detection and restoration of contamination and impurities in spectral images (python)

HydrogenPB – a code written in Wolfram Mathematica that calculates the Stokes I and V profiles of a hydrogen atom emission under arbitrary magnetic fields. The code takes into account the fine structure of hydrogen atom terms, the intermediate Paschen-Back effect. It is now in MVP stage, and is also rewritten in python as a part of SolRaT.

AdaptFMin – a code for an effective numeric solution of functional minimization functional using finite difference and finite element methods with on-the-fly adapting mesh. This code is written in python, is now in MVP stage, and is actively used as one of the methods for finding the equilibrium liquid crystal director profiles.