Skip to content
View yezdata's full-sized avatar
3 followers · 1 following

Highlights

  • Pro

Block or report yezdata

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
yezdata/README.md

Welcome to my GitHub! ⚡

  • 🤖 I'm passionate about machine learning (particularly NLP, DeepLearning and uncoventional ML architectures), statistics, programming and technology.
  • 🌱 I’m currently studying Mathematical and Statistical Methods in Economics at Masaryk University.
  • 🔭 Now working on these projects:

    • as ML / NLP Research Engineer: External contractor and member of research grant team for the project: New Methods for Emotion-based Economic Uncertainty Modelling.
      Project focus & technical scope:

      • Developing and fine-tuning Transformer-based models for emotion recognition and sentiment classification in economic discourse.
      • Applying NLP techniques to article news data to enhance the Economic Policy Uncertainty (EPU) index.
      • Designing end-to-end processing and inference pipeline for EPU index.

    • PyDoctor, which is CLI utility providing docstring insertion into python codebase. Python files are first parsed using the libCST library to build target code with relevant context. PyDoctor then uses local custom finetuned Instruct LLM to generate docstrings and then atomically writes the file back with the changes. The CLI also contains standard QoL options such as per file ignore, per code block ignore, --dry-run (with difflib logs), change the target code blocks (with docstring, without, all)...

How to reach me 📫

Email jez.vasek@gmail.com
LinkedIn https://www.linkedin.com/in/vaclavjez

Tech stack 🛠️

Python R SQL Git Docker Linux JavaScript HTML CSS

Pinned Loading

  1. pydoctor pydoctor Public

    🩺 A doctor for your Python docstrings. Fully local CLI that auto-generates docstrings with a fine-tuned LLM, documenting your codebase.

    Python 5

  2. emcoder emcoder Public

    DeepLearning Probabilistic Emotion Recognition model with MC Dropout Uncertainty Quantification. Lightweight architecture, transparent reliability.

  3. pydocs_ai pydocs_ai Public

    📚 Chat with your Docs: RAG-based assistant to provide hallucination-free answers via semantic chunking, indexing and retrieval of Python libraries documentation.

    Python 1

  4. bayes_vs_ml bayes_vs_ml Public

    ⚖️ Glass Box vs. Black Box: comparative study of interpretability vs. performance. Contrasts Bayesian Inference (PyMC) against Fine-Tuned Transformers (FinBERT) on financial sentiment classification.

    Python

  5. pyconv pyconv Public

    🎧 Local-first async audio pipeline for speech detection, Faster-Whisper transcription, and real-time conversation classification.

    Python

  6. financial_volatility_forecaster financial_volatility_forecaster Public

    📊 Predict market risk with precision. A full-stack volatility engine featuring real-time GARCH inference, automated Nasdaq-100 evaluation and performance tracking.

    Python 1