I am Aleks Mashanski, a Documentation Developer who is interested in the framework of Cybersecurity and AI. I also have an experience in Coding Practice Challenges and Data Science field.

A lot of my repositories & pet projects represented here, from my school days when I tried Java and C languages to nowadays, where mostly of my projects are related with Python and AI in general. I'm in relations with Further Mathematics since my secondary school. And I really like the fact that things like Mathematics, Combinatorics, Statistics & Probability Theory can be combined with Programming and generate an interesting result!

In addition, I’m interested in the mathematical aspects of Cryptography, various Cryptographyc Algorithms & Methods.

• Languages: Python, JavaScript, SQL
• Data Formats: JSON, XML

• Authoring Tools: MadCap Flare, Markdown
• Static Site Generators: Docusaurus
• Methodologies: Docs-as-Code
• Markup Languages: XML, HTML

• Web Technologies: RESTful APIs (JSON and XML)
• Runtime: Node.js

• Operating Systems: Linux (Debian, Ubuntu) + Shell Scripting
• Containerization: Docker
• Version Control: Git (GitHub, GitLab)

• Language Models & NLP: Solid understanding of modern LLMs (e.g. GPT, BERT, T5), Word2Vec, contextual embeddings, fine-tuning & prompt engineering
• Supervised Learning: Linear & Logistic Regression, Decision Trees, Random Forests, Gradient Boosting (XGBoost, LightGBM), K-Nearest Neighbors, Naive Bayes
• Unsupervised Learning: Clustering (K-Means, DBSCAN), PCA, t-SNE, anomaly detection
• Model Evaluation: Cross-validation, ROC AUC, F1-score, hyperparameter tuning (Grid, Random, Optuna)

• Python stack: NumPy, Pandas, Scikit-Learn
• Boosting frameworks: XGBoost, LightGBM, CatBoost
• Automated ML: LightAutoML, H2O.ai, Vowpal Wabbit
• Deep Learning: PyTorch (preferred), TensorFlow/Keras
• Computer Vision: OpenCV, TorchVision, basic experience with Transformers for Vision
• Visualization: Matplotlib, Seaborn, Plotly (for interactive dashboards)

• Data Pipeline: Data Cleaning, Preparation, Feature Engineering & Selection
• Time Series: Forecasting, seasonality, ARIMA, Prophet
• Natural Language Processing: Tokenization, vectorization, sentiment analysis, text classification, embeddings
• Modern NLP Practices: HuggingFace Transformers, zero/few-shot learning, text generation
• Neural Network Architectures:
  • CNN – for image classification & feature extraction
  • RNN / LSTM – for sequential data
  • Transformer-based – foundational knowledge of self-attention and encoder-decoder structures
  • VAE – for generative modeling and anomaly detection
  • GAN – for synthetic data generation

• English (Fluent)
• Polish (Fluent)
• Russian (Native)
• German (Basic)
• Georgian (Basic)

• Leadership & Ownership
• Cross-functional Team Collaboration
• Time & Priority Management
• Self-Motivation & Accountability
• Clear & Effective Communication
• Adaptability in Fast-Paced Environments

• Exploring Cybersecurity and AI learning material
• Crunching Codewars

• Open Source projects related to AI, Computer Vision and LLMs

• Linkedin: https://www.linkedin.com/in/maszanski/
• Website: https://mashanski.me
• Kaggle: https://www.kaggle.com/alexmaszanski
• Mail: [email protected]
• Twitter: https://twitter.com/maszanski

My articles explore both core data science concepts and hands-on machine learning methods, from tutorials to algorithm deep-dives.

• Probability theory for beginners

• Using NumPy for Python matrices

• CDN (Content Delivery Network)

• What is reinforcement learning and how it works

• Naive Bayesian classification algorithm

• The k-nearest neighbour method

• Machine Learning for Beginners: The Random Forest Algorithm

• Solving machine learning problems with a gradient boosting algorithm

• Feature engineering and selection. Part 1

• Feature construction and selection. Part 2

• Kaggle in 30 minutes: a practical guide for beginners. Part 1

• Kaggle in 30 minutes: House Prices competition. Part 2