Senior Data Scientist, Spark, Feature Engineering

MGID

full-time

Posted on: 9/25/2025

Origin: • 🇺🇦 Ukraine

✨ AI Apply

Senior

HadoopPythonPyTorchScikit-LearnSparkSQLTensorflow

About the role

Join MGID as a Data Scientist and become part of a team shaping the future of AdTech.
Explore data to identify trends, evaluate campaign performance, and support decision-making.
Design, implement, and optimize algorithms for CPA (Cost Per Action) optimization.
Build and refine predictive models to improve decisions for CPA campaigns, ensuring alignment with business objectives.
Work closely with product and business teams to translate technical solutions into measurable business impact.
Conduct and evaluate A/B tests to assess the impact of algorithm changes.
Create dashboards and visualizations to track key performance indicators (KPIs) and communicate findings to stakeholders.
Extract maximum value from data to inform product development and increase conversion volume and budget efficiency.

Proactive problem-solving attitude: ability to independently analyze data, identify hidden issues, formulate hypotheses, suggest effective solutions, develop and test them.
Ability to analyze and deeply explore data to extract valuable insights; ability to visualize and communicate those insights.
Proven experience in data science, with a strong understanding of optimization problems.
Proficiency in Python and SQL, including data manipulation and feature engineering.
Expertise in machine learning techniques and tools (e.g., CatBoost, XGBoost, TensorFlow, PyTorch, scikit-learn).
English - Upper-Intermediate.
Will be a plus: MSc or PhD in Engineering, Statistics, Mathematics or related fields.
Will be a plus: Familiarity with the AdTech domain.
Will be a plus: Experience with big data technologies (Spark, Hadoop) and processing large datasets.
Will be a plus: Hands-on experience with A/B testing and business metric optimization.
Will be a plus: Background in working with high-frequency or large-scale event data (e.g., 15B+ events daily).