Aleksandra Kiuberis

Project producing high-quality analytical data with quantified uncertainty for strategic decision-making. Outputs directly shaped prioritization of high-cost technical initiatives.

• • Designed and maintained automated ETL data pipelines in Python, converting raw computational outputs into structured, analytics-ready datasets. Reducing reporting setup time from hours to minutes.

• • Deployed scalable datasets for international partners to support predictive modeling.

• • Developed Python/SQL data quality checks, reducing validation time from weeks to days.

• • Coordinated reporting workflows across 9+ international teams.

• • Created dashboards and visual reports to support policy and operations.

Project involved transforming limited computational data into millions of high-accuracy predictions using ML-based surface modeling. Results accelerated simulation pipelines and improved decision-making around target systems.

• • Designed predictive pipeline (100k+ entries) from 200-300 simulations.

• • Built preprocessing, feature engineering, and validation workflows in Python.

• • Applied Gaussian Process Regression to generate high-accuracy predictions and replace simulation-heavy workflows.

• • Reduced compute cost and turnaround time by 90% compared to direct simulations.

• • Applied model validation and monitoring for scalable.

Project delivering validated, high-quality datasets for global modeling tools. Outputs improved data quality for downstream climate and astrophysics workflows.

• • Prepared and delivered multi-million-record datasets to global repositories (HITRAN, GEISA, EXOMOL).

• • Developed internal Python scripts for automated data preparation and validation, processing hundreds to thousands of input files per dataset.

• • Built reproducible preprocessing workflows to replace manual data handling and reduce human error.

• • Standardized QA across international partners (UK, USA, Germany), reducing error rates by 20%.

• • Authored internal technical documentation and trained team members to use data processing tools and workflows.