Taras S. — Data Science Engineer

Work Experience

Data Scientist (Computer Vision)

Duration: 1 year 6 months
Summary: 

• Researched and implemented pipeline for synthetic image transformation using noise generation by simulating noise characteristics from cameras. Implemented two approaches parametric and non-parametric based on statistical analysis.

• Development of a system for reprojection of 3D dog models based on images to display animal animations on a device. Utilizing an architecture based on PyTorch and a custom-collected dataset. The process involves obtaining keypoints and segmentation from heatmaps, processing statistical components, calculating rotations, and generating the 3D dog model.

Responsibilities:

• Rewrite the code from PyTorch implementation to Keras

• Conduct separate parts of whole architecture of the Barc

• Ensure compatibility of a new model with custom dataset

• Train pretrained models based on in-house dataset

• Validation of in-house dataset

• Train and evaluation of all pipeline model

• Visualisation of both final and intermediate results of the pipeline model

Data Scientist, Logistics Route Optimization

Duration: 6 months
Summary: Completed data engineering to improve quality of data (such as creating for filling missing data, etc). Exploratory data analysis and machine learning solutions aimed to determine the most optimal route considering port characteristics. It includes statistical models, regression models, time series analysis, recommendations, and various custom approaches.

Responsibilities:

• Used SQL-like databases for data engineering as well as data extraction
• Executed the EDA of port-related data
• Analyzed data for the insights and patterns, as well as presented results of investigation
• Researched the approaches to visualize the data insights for customers
• Implemented ML solutions for estimating the port cost
• Used SQL-like databases for data engineering as well as data extraction
• Executed the EDA of port-related data
• Analyzed data for the insights and patterns, as well as presented results of investigation
• Researched the approaches to visualize the data insights for customers
• Implemented ML solutions for estimating the port cost

Technologies: Python, Pandas, SciPy, Scikit-learn, Matplotlib, Plotly, Prophet, BigQuery

Data Scientist, Search Engine Optimization (SEO)

Duration: 1 year
Summary: Development of custom solutions for clustering keywords by semantic similarity and other characteristics, including web application for a demo based on Django.
Technologies: Python, Pandas, Numpy, Scikit-learn, Matplotlib, spaCy, USE, GPT3, Django, Docker, Kubernetes, AWS S3

Education

• Bachelor’s Degree in Mechanical Engineering
• Master’s Degree in Computer Engineering, ongoing

Certification

• Kyivstar Big Data School 4.0
  Program of the school introduces students to major concepts and techniques of data science process: predictive analytics and machine learning at scale, big data tools and technologies, basics of business analytics.
  2019
• Data Science Camp
  Offline ML course at SmartInsight.
  2021
• Sequences, Time Series and Prediction
  Coursera course.
  2023
• Convolutional Neural Networks in TensorFlow
  Coursera course.
  2023
• Introduction to TensorFlow for Artificial Intelligence, Machine Learning, and Deep Learning
  Coursera course.
  2023
• Natural Language Processing in TensorFlow
  Coursera course.
  2023