Lundrim Alla

I’m a passionate and detail-oriented Data Science student with a strong foundation in computer science, statistics, and machine learning. Currently pursuing my master’s degree at the University of Vienna, I focus on extracting insights from complex data, especially in the areas of text analytics, clustering, and data visualization.

My academic journey has given me hands-on experience in working with real-world datasets across multiple domains, from customer behavior to audit findings. I enjoy working on problems that involve transforming raw, unstructured data into meaningful representations. I have applied a range of techniques such as sentence embeddings (BERT, GPT-2), dimensionality reduction (UMAP, PCA, t-SNE), and clustering algorithms (K-Means, Agglomerative Clustering, DBSCAN), always with a focus on interpretability and evaluation using internal and external metrics like Silhouette Score and Adjusted Rand Index (ARI).

I’m also interested in improving the way we evaluate clustering results. Recently, I’ve been exploring advanced validation methods such as DBCV and fairness-aware clustering, including techniques that ensure individual and group-level fairness in data grouping. My background includes experience with topic modeling (LDA), text preprocessing, and visualization techniques such as word clouds, heatmaps, and interactive dashboards.

From a technical perspective, I’m highly proficient in Python, using libraries like pandas, scikit-learn, matplotlib, seaborn, transformers, and umap-learn. I’ve also worked with SQL, Tableau, and D3.js for front-end visualizations, and I'm comfortable managing environments with tools like Jupyter Notebook, VS Code, and Git.

Beyond technical skills, I value clarity, clean code, and structured experimentation. I document my work carefully and always aim for reproducibility and explainability, especially when working on unsupervised learning tasks.

Whether it’s for academic research, a side project, or a freelance collaboration, I’m always excited to work on data-driven problems that require analytical thinking, creativity, and practical implementation.