At the International Air Transport Association (IATA), I implemented statistical anomaly detection methods to identify outliers across large, multi-domain datasets. I validated findings against external news articles and official statistics to establish a baseline for model evaluation.
I developed and tested machine learning anomaly detection models, including Isolation Forest, DBSCAN, Local Outlier Factor, autoencoders, and the Darts library. I also built interactive Power BI dashboards showing estimated and forecasted traffic trends and competitor statistics across routes and segments.
On a five-person research project, I contributed to fine-tuning a large language model for zero-shot reasoning in natural language processing applications. In a separate music analysis project, I cleaned and transformed data and used Python visualizations to explore changes in popular songs over time.

