Alin Airinei


11743


Email: alin.airinei924@gmail.com              LinkedIn: https://www.linkedin.com/in/alin-airinei/
            
GitHub: https://github.com/Al-1n             Portfolio: https://al-1n.github.io/portfolio/


Summary


Physics graduate with a passion for data analysis and a strong desire to contribute to meaningful projects. Proficient in utilizing statistical analysis techniques and modern data visualization tools to derive actionable insights and solve complex problems.


Education


Bachelor of Science in Physics, Stony Brook University, 2021

                       •            

IBM Data Science Professional Certificate, Coursera


Portfolio Projects


1. Optimizing Gradient Boosting Models for Medical Diagnosis

• Compared the performance gains from hyperparameter and threshold tuning with Hyperopt, Optuna, and TunedThresoldClassifierCV on Gradient Boosting.

• Improved recall in diabetes diagnosis from 0.64 to 0.89 using gradient boosting models with hyperparameter and threshold optimization.

• Achieved up to 15% increase in F1-scores across XGBoost, LightGBM, and CatBoost models through combined optimization techniques.

• Enhanced AUPRC scores by up to 10.7% for XGBoost and LightGBM models, indicating significant improvement in precision-recall trade-offs for imbalanced data.

• Attained consistent AUROC scores of 0.83 for all three models post-optimization, demonstrating robust classifier performance.

• Technology Used: Python, Streamlit, Plotly, Pandas, NumPy, Geopy, Jupyter Lab.

Full Report | Project Page

2 . SpaceX Launch Records

• Utilized data science tools to dissect SpaceX’s successful business model.

• Comprehensive implementation of typical data analysis and machine learning processes.

• Technology Used: Python, SQL, Pandas, Plotly Dash, Beautiful Soup, Scikit-learn, Folium.

Full Report | Project Page

3. Superconductivity and Quantum Computing

• Analyzed data from a real-world quantum mechanics experiment involving Josephson junctions.

• Tested and validated the alignment of theoretical predictions against experimental findings.

• Technology Used: Python, Jupyter Lab, Pandas, NumPy, Lmfit, Matplotlib.

Project Page | Scientific Report


Skills


Technologies:

  • Programming Languages: Python, SQL
  • Data Visualization Tools: Matplotlib, Seaborn, Plotly, Streamlit, Tableau
  • Databases: MySQL, PostgreSQL
  • Version Control: Git
  • Data Modeling: Machine Learning, Lmfit

Techniques:

  • Statistical Analysis: Descriptive Statistics, Regression Analysis, Distributions, Correlation Analysis
  • Data Manipulation and Cleaning: Imputation Techniques, Feature Engineering
  • Data Visualization: Plot Customization, Interactive Dashboards, Maps