Віталій

VITALII DAINEKO
Junior Data Scientist
[открыть контакты](см. выше в блоке «контактная информация») · [открыть контакты](см. выше в блоке «контактная информация»)
[открыть контакты](см. выше в блоке «контактная информация») · https://github.com/data-tamer2410 · Aix-les-Bains, France

Summary
Results-driven Data Scientist with expertise in LLMs, Transformers, and end-to-end machine learning.
Skilled in data analysis, NLP, web scraping, API development, and model fine-tuning for production. Strong
background in data integration, database management, and interactive dashboard creation. Passionate
about solving complex problems and building scalable solutions.

Hard Skills
Programming & Libraries: Python, SQL, MongoDB, TensorFlow, Keras, Scikit-Learn, Transformers, Numpy,
Pandas, Matplotlib, Seaborn, BeautifulSoup, Streamlit, FastAPI, NLTK, Requests, Surprise, Optuna.
Tools: PowerBI, Docker, Poetry, Git, Conda, Bash.

Soft Skills
Team collaboration, strong communication, critical and analytical thinking, problem-solving, adaptability to
new tools and technologies.

Project Experience
Doctor Chat – AI Assistant for Medical Consultations
https://github.com/data-tamer2410/ds-doctor-chat
Tech Stack: TensorFlow, Transformers, Docker, Sreamlit, Pandas.
In this project, I developed an AI-powered chatbot that provides medical advice and recommendations.
The bot utilizes a fine-tuned GPT-2 model to generate responses in a structured dialogue format between
the patient and a virtual doctor. The application is built using Streamlit and deployed on a server, ensuring
seamless accessibility. The chatbot offers personalized health advice, making it an essential tool for quick
consultations and initial diagnostics.
Achievements: Fine-tuned the GPT-2 model to achieve high accuracy in generating medical advice,
reducing perplexity to 12.91. Implemented a rigorous fine-tuning process to tailor the responses specifically
to the medical domain, ensuring that the chatbot provides accurate, reliable, and contextually appropriate
answers. Built an intuitive and engaging front-end with Streamlit, allowing users to interact easily with the
bot and receive personalized advice.
Weather Forecast API
https://github.com/data-tamer2410/ds-weather-forecast
Tech Stack: Keras, FastAPI, Docker, Numpy, Pandas, Scikit-Learn, Matplotlib.
In this project, I developed an RNN-based API to forecast weather conditions for the next day using
historical data from the past 7 days. The API was containerized using Docker and deployed on a server,
enabling seamless integration and scalability. The endpoint is easily accessible via POST requests, making it
practical for real-world applications in weather monitoring or decision support systems.
Achievements: Designed and deployed an RNN model optimized for time-series weather data prediction.
Built a FastAPI backend, packaged it in Docker, and deployed it on Render. Developed an interface that
accepts 7 days of weather data and returns precise forecasts, including the probability of rainfall.
Implemented scaling and reshaping mechanisms to prepare complex weather datasets for accurate model
inference.

Education
Separate Structural Subdivision "Kryvyi Rih Vocational College of the National Aviation University" UA
Associate's Degree in Computer and Software Engineering 2022 - 2025
IT School GoIT, Data Science and Machine Learning Aug 2024 - Dec 2024

IT School GoIT, Python Core May 2024 - Aug 2024

Languages
English: Pre-Intermediate (improving) Ukrainian: Native Russian: Native