Skip to content
View theMagusDev's full-sized avatar

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Highlights

  • Pro

Organizations

@ConfereeBot @Project-Codular

Block or report theMagusDev

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
theMagusDev/README.md

Yuriy Magus

ML/NLP Engineer

I am an ML Engineer focused on NLP, multi-agent systems, and classic ML. I prioritize projects that deliver a direct, measurable impact on business growth and user experience.

About me

  • Business-Centric: I build AI solutions to solve real-world problems, not just for the sake of technology.
  • Reliability: strict follower of the GTD (Getting Things Done) system — I never miss a deadline.
  • Team Player: I thrive in collaborative environments and enjoy achieving ambitious goals together.

Tech Stack

ML & Data Science

  • NLP/LLM: LangChain, LangGraph, Unsloth (QLoRA), Transformers, ChromaDB, LangSmith.
  • Classic ML: Scikit-learn, CatBoost, XGBoost, LightGBM, PyTorch, MLflow.
  • Data: SQL, NumPy, Pandas, Matplotlib, Seaborn, SciPy.

Auxiliary tools (Backend & DevOps)

  • Languages: Python (FastAPI, Pydantic), Go, Java/Kotlin.
  • Infrastructure: Docker, Git, Bash.

Core Competencies

  • LLM Fine-tuning: experience in Knowledge Distillation (Teacher-Student) and model optimization.
  • RAG Systems: building end-to-end retrieval pipelines with advanced chunking and vector storage.
  • Agentic Workflows: developing summarization and translation agents using multi-agent architectures.

Pinned Loading

  1. RAG-system RAG-system Public

    Retrieval-Augmented Generation (RAG) system using LangChain, tiktoken, langchain-openai, and ChromaDB for document ingestion, embedding, vector search, and LLM-powered querying

    Jupyter Notebook

  2. fine-tuned-description-enhancer fine-tuned-description-enhancer Public

    Fine-tuned Qwen 2.5 7B for professional ad copywriting. Built with Knowledge Distillation (DeepSeek V3) and QLoRA (Unsloth). Improves UGC structure and selling style. Optimized for T4 GPU.

    Jupyter Notebook

  3. Dncoder-Decoder-RNN-for-NMT Dncoder-Decoder-RNN-for-NMT Public

    Seq2Seq NMT (PyTorch): Implementation of the classic Encoder-Decoder RNN architecture for Neural Machine Translation (German to English) featuring a modular project structure and using the Multi30k…

    Jupyter Notebook

  4. names-generator-RNN names-generator-RNN Public

    Implementation of a Char-RNN with PyTorch for sequence generation (names). Features a modular structure, temperature sampling for inference, and character-level tokenization.

    Jupyter Notebook

  5. credit-scoring credit-scoring Public

    Credit scoring system pipeline with XGBoost, SHAP interpretability, MLflow tracking and statistical significance testing on the "Give Me Some Credit" dataset.

    Jupyter Notebook

  6. retailrocket-recommender-system retailrocket-recommender-system Public

    Recommender system for e-commerce

    Jupyter Notebook