Skip to content

A portfolio of various past projects in deep learning and time series analysis.

Notifications You must be signed in to change notification settings

Rokcernejr/Rok-Cerne-Portfolio

Repository files navigation

Machine Learning Engineering Portfolio

About Me

Machine Learning Engineer specializing in deep learning and quantitative analysis, with expertise in building containerized ML pipelines and predictive models. Passionate about combining supervised and unsupervised approaches to solve complex real-world problems. Former materials scientist who brings a unique perspective to data architecture and model optimization.

🚀 Featured Projects

Containerized Document Processing Pipeline

  • Developed an innovative ML pipeline combining supervised and semi-supervised learning
  • Reduced manual labeling requirements by 90% through semi-supervised techniques
  • Built with TensorFlow, Flask, Docker, and PDF processing capabilities
  • Implemented pseudo-labeling and consistency regularization for efficient data handling
  • Tech Stack: Python, TensorFlow, Flask, Docker, pandas, NumPy, scikit-learn, Gunicorn

NLP Spoiler Detection Model

  • Implemented three approaches: Logistic Regression, BiLSTM with GloVe, and ALBERT
  • Conducted comprehensive EDA on IMDB movie reviews
  • Optimized performance through sophisticated feature engineering
  • Analyzed genre-specific patterns and review characteristics
  • Tech Stack: Python, PyTorch, Transformers, scikit-learn

Apple Trading Strategy Analyzer

  • Combined sentiment analysis from Bloomberg News with 1D-CNN features
  • Developed comparative models using traditional financial indicators
  • Implemented comprehensive validation methods and performance metrics
  • Created detailed visualizations for model interpretation
  • Tech Stack: Python, TensorFlow, NumPy, pandas, Sentiment Analysis tools

Social Network Cluster Analysis

  • Analyzed Facebook network data (4039 nodes, 88,234 edges)
  • Identified distinct user clusters using various algorithms
  • Engineered network-specific features for improved analysis
  • Achieved 0.59 silhouette score with Agglomerative Clustering
  • Tech Stack: Python, NetworkX, scikit-learn, UMAP

Fine-Tuned RNN for Market Prediction

  • Enhanced existing market prediction model with improved feature engineering
  • Optimized model parameters through cross-validation
  • Implemented robust validation methodology
  • Developed sophisticated risk management mechanisms
  • Tech Stack: Python, TensorFlow, Technical Analysis tools

🛠 Technical Skills

Languages & Frameworks

  • Python (Primary)
  • TensorFlow
  • PyTorch
  • Flask
  • scikit-learn

Tools & Technologies

  • Docker
  • Git
  • CI/CD
  • Google Cloud Platform
  • Jupyter

Machine Learning Specialties

  • Deep Learning
  • Natural Language Processing
  • Time Series Analysis
  • Computer Vision
  • Semi-Supervised Learning
  • Reinforcement Learning

📫 Get in Touch

Feel free to reach out for collaborations or just a chat about machine learning!

Email: [email protected]


This portfolio is continuously updated with new projects and improvements.

About

A portfolio of various past projects in deep learning and time series analysis.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages