At Defined.ai, I design data quality controls for enterprise pipelines, translating client and business requirements into testable standards, thresholds, and escalation paths. A reusable Python/SQL validation layer and SQLite-backed sampling system I designed reduced post-delivery defects by 35%.
I also design and maintain data clean-up and standardisation processes across 30+ locales, using Python, Unicode normalisation, regex, and Azure OpenAI. I partner with product and engineering teams to build governance into the platform and provide quality reporting through Power BI dashboards and SQL Server-backed reports.
Before Defined.ai, I managed longitudinal research datasets at Phonetics and Phonology and Lisbon Baby Lab, from collection and curation through analysis and archiving. My research in Linguistics grounded my work in experimental design, sampling, inter-rater reliability, and statistical inference.
