During my research internship with the Sydney Institute for Astronomy at the University of Sydney, I worked on data-driven analysis related to early-Universe supermassive black holes.
The practical challenge was not only running an analysis. It was making the complete process reproducible: loading and cleaning datasets, recording assumptions, visualizing distributions, estimating parameters, and preserving enough context for another researcher to understand the result.
Tools and methods
- Python and Jupyter for repeatable analysis notebooks.
- Pandas and NumPy for data preparation and numerical work.
- SciPy for statistical parameter estimation.
- Astronomy-focused scientific tooling for domain-specific workflows.
- Clear visualizations to communicate uncertainty and trends.
I also supported structured case-study work around compact objects and black-hole astronomy, including Sagittarius A*. The experience strengthened my interest in research software: code that is not merely correct once, but transparent, repeatable, and useful to other people.
That mindset now carries into my product engineering work, especially when building AI systems that need evidence, evaluation, and reliable data pipelines.