Fișier CV încărcat

Passion and Goals

I have found my passion in Data Science and AI. I am eager to apply innovative ideas into practice and contribute significant development in the field of AI, which keeps evolving daily and brings a lot of innovation into our daily life. My goal is to continuously learn and improve my skills and work with driven and inspiring people. I would love to contribute to team projects that create real impact.

Projects

• Bachelor Thesis | Minimising false positives in Intrusion Detection Systems with Machine Learning
  Feb 2025 – June 2025

  • Analysed 148k network connections from the NSL-KDD benchmark (train = 125,973, test = 22,544) to classify attack vs normal traffic.
  • Benchmarked the performance of 5 supervised learning methods, including XGBoost, Naïve Bayes, and Logistic Regression, to minimise false positives while preserving a high recall.
  • Engineered categorical features with One-Hot Encoding and automated hyperparameter tuning (GridSearchCV and 5-fold CV) to maximise F1-score.
  • Cut the false positive rate to 2.6% while retaining 97% precision and 65% recall on the held-out test set.
  • Created a hand-made rule-based baseline classifier analysing each feature separately and validated it against XAI SHAP values to refine alert logic.

• Group Project 3 | Healthcare Cost Prediction using Machine Learning and XAI
  Feb 2024 - Jun 2024

  • Modelled annual premiums for 15,000 US individuals using demographics, lifestyle, and clinical data.
  • Benchmarked five regression models with tree ensemble methods achieving >10x lower MAPE than ANN baseline, keeping error lower than 10%.
  • Conducted an in-depth study on Explainable AI, focusing on intrinsically interpretable models and post-hoc explanation methods, including Partial Dependence Plots (PDPs) and local interpretability techniques.
  • Evaluated models' robustness by analysing sensitivity to noise through systematic input perturbations, identifying weaknesses, and improving interpretability in decision-making processes.

• Group Project 2 | Breakthru Board Game
  Sep 2023 - Jan 2024

  • Implemented a Java-based Breakthru board game application.
  • Developed a MiniMax and Monte Carlo Tree Search (MCTS) agent with adaptive learning capabilities.

• Group Project 1 | Titan Space Mission Simulator
  Feb 2023 - Jun 2023

  • Created a Java-based planetary motion simulation using differential equation solvers, mathematical modeling, and physics concepts to produce precise trajectory computations.
  • Designed an interactive GUI to visualize the movement of the solar system and the probe's path.

Education

• Bachelor of Science (BSc) Data Science & Artificial Intelligence
  Sep 2022 - Jun 2025
  • Maastricht University, Maastricht, The Netherlands
  • Machine learning | Mathematics & Statistics | Natural Language Processing | Data Analysis | Engineering.

Skills & Certifications

• Technical Skills: Java, SQL, MatLab, Jupyter Notebook, Python (NumPy, TensorFlow, Scikit-learn).
• Other Skills: Team-player, adaptability, accountability, ability to work independently, critical thinker.
• Languages: English, Russian, Romanian - Fluent.