Особисті дані приховані

Daria Iskusnykh
Graduate student in Data Science and Machine Learning with basic experience in
data analysis, neural networks, and visualisation. I learn quickly, have analytical
thinking skills, and pay attention to detail.

[відкрити контакти](див. вище в блоці «контактна інформація»)
[відкрити контакти](див. вище в блоці «контактна інформація»)
Linz, Austria

TECHNICAL SKILLS PROJECTS

Languages: Python (scikit-learn, PyTorch, Audio Onset Detection Using Deep Learning
TensorFlow), SQL, R Python, PyTorch, NumPy, SciPy, Conv1D, BiLSTM, CRNN, Signal
Data & EDA: pandas, NumPy, matplotlib, seaborn Processing, Deep Learning
Visualisation: Power BI, Excel (pivot tables),
dashboards Developed a neural network for musical onset detection using a
Tools: Git, Jupyter, Google Colab, REST APIs hybrid CNN–BiLSTM architecture and handcrafted signal features
Cloud & APIs: Google Cloud (Vision API), OpenAI API, (Spectral Flux, RMS, ZCR, etc.).
API querying Also implemented beat and tempo detection modules using CRNN +
Concepts: Classification & regression, Model Attention and LightGBM ensemble
evaluation, Clustering, Feature engineering & Solo project – 3rd best result among all teams.
preprocessing, Deep learning, Data bias & fairness,
EDA and hypothesis-driven analysis Bachelor Thesis "Information System for Analyzing the Quality of
Object Recognition in Pictures"
Python, Pandas, NumPy, Jupyter Notebook, YOLOv8, Google Cloud
LANGUAGE Vision API, Clarifai

Assessing the accuracy of neural networks in image recognition,
ENGLISH - B2 (IELTS 2020) focusing on algorithms' resistance to CAPTCHA. Comparing several
GERMAN - A2 open-access algorithms. Investigating methods of distorting
UKRAINIAN, RUSSIAN - NATIVE images, making them difficult for neural networks but readable for
humans. Assessing the effectiveness of human vision vs. image
recognition algorithms.
EDUCATION
Investigating bias of recommended systems datasets
Python, Pandas, NumPy, Decision Tree Classifier, Bias Detection,
Johannes Kepler Universität Linz (JKU Linz) Data Analysis
Bachelor Artificial Intelligence
2021 - now Several datasets from open WEB sources and/or web scraping with
user interaction data were analyzed to identify bias in
Vasyl' Stus Donetsk National University (DonNU) recommendations. Key types of bias in data were identified
Bachelor Computer Science (Position bias, Exposure bias, Popularity bias). Analysis of the
2020 - 2024 influence of element popularity on the results of recommendations
and methods for solving the problem was considered.
CERTIFICATES
Custom GPT AI Assistant
Python, OpenAI GPT, Prompt Engineering, AI in Education
Operations Research: Optimization Algorithms
Coursera, National Taiwan University Explored the application of AI in teaching. Developed a custom GPT
assistant for teaching a specific subject using Prompt Engineering.
IT-products marketing Configured a response system to ensure academic integrity and
EdEra, Genesis IT personalized learning.

Creation and development of IT-products Power BI Data Visualization & SQL Analysis
EdEra, Genesis IT Power BI, SQL (MySQL, PostgreSQL), Data Analysis

Conceptual & Semantic Data Modeling Experience with Power BI includes creating interactive reports,
JKU Linz, Area of specialization processing data in Power Query, and using DAX calculations to
analyze trends and sales. Data models and relationships between
tables were developed.