Stepan

Stepan Sushko

Data scientist

Education
Bachelor of Computer Science
From September 2020 to May 2024 Ukrainian Catholic University Lviv
• My team took third place in hackathon by UCU
• Several times managed team consisting of up to 5 people for university projects

 [открыть контакты](см. выше в блоке «контактная информация») Work experience
 [открыть контакты](см. выше в блоке «контактная информация») Machine Learning Developer
 https://github.com/stepansus Since January 2023 Logivations Lviv
hko1
• Developing accurate real-time 3D pose tracking of challenging entities
• Deploying models to embedded devices with maximum efficiency, using modern
tools such as nvidia-docker, DeepStream and TensorRT
• Used Gazebo simulation for object tracking and simulation of real camera view
• Deploying models functionality to company's website via fastapi
Data Scientist
Languages From March 2022 to September 2022 MountGuide Lviv

Ukrainian • Took part in building and improving back-end part for app
• Annotate and create datasets from scratch
Fluent
• Build feature for finding name of mountain uploaded by user from photo
English
Advanced
German Independent Projects
Intermediate (B1 Goethe
Zertifikat) Research. 3D Reconstuction from several images
Currently I take part in research at my university.
• generation of a 3D model from a collection of 2D images, captured from different
viewpoints.
Skills • Recently, we have been exploring the use of neural radiance fields models for 3D
reconstruction.
Python (numpy, scipy, pandas,
scikitlearn, fastapi, Speech Emotion Recognition Model
tensortflow, keras, cv2) • Implemented Support Vector Machine for speech emotion recognition.
MySQL, MSSQL, Cassandra • Sequential Minimal Optimization algorithm was used as a solver
• Mel-frequency cepstral coefficients were used to work with sound.
C++ (multiprocessing, • Methods and model written from scratch.
multithreading)
Electric door lock using FaceID on Raspberry
Git. docker
• Designed the electrical door lock which was managed by the app
built with Android Studio.
• The owner can view event logs, check photos of the intruder, view and create users.
• Neural network was used for face recognition to open locker.
• python, tensorflow, cv2, java, Firebase
Quad-Drone on Raspberry
• Сonstructed and programed a quadcopter
• Drone could automatically fly over a given area, take photos and generate a 2D map
• python, paramiko, uart, embedded

Courses
Machine Learning (Supervised, unsupervised, deep learning)
Math for Machine Learning (optimization, probability, linear algebra methods)
Big Data (principles of building distributed big data processing systems)
Forecasting (forecasting theory, classical, statistical approaches, ML, DL techniques,
graph theory)
Database Organisation (Relational DB)
Linear Algebra (Matrices, linear systems, spaces and transformations, orthogonality,
eigenvalues)
Probability & Statistics (Random variable, Limit theorems, Parameter estimation,
Hypothesis testing)
Algorithms & DS(sorting, heap, binary tree, hashing, greedy algorithms, dynamic
programming, graph)
Calculus (mathematical analysis of applied problems)