Darwin, NT, Australia

Built a not-for-profit's first analytics infrastructure, moving four disconnected systems into one cloud database serving an ACNC-registered community centre.

• Built the organization's first data pipeline, moving registration, attendance, feedback and finance data from four disconnected systems into one cloud database with Python and SQL, replacing fully manual reporting.
• Reconciled member identities across class attendance books and systems as students moved between classes, creating one trusted member record for downstream reporting.
• Designed the warehouse layer structure, keeping raw landing tables separate from the cleaned, modeled tables analysts query.
• Built executive Power BI dashboards with near real-time views of engagement and retention across 800+ members.
• Surfaced emerging participation and funding gaps while cutting manual reporting effort by 40%, giving leadership faster access to the numbers behind decisions.
• Worked within a volunteer budget, choosing low-cost tools while still delivering a system the organization could actually run.

Melbourne, VIC, Australia

Collected 2026 Census data face to face for people less likely to use the standard online process, under strict privacy and data-quality procedures.

• Collected 2026 Census data face to face for groups less likely to use the standard online process, including people in hospitals, aged care homes and homeless shelters.
• Ran public pop-up hubs and information sessions, then worked one on one with people to complete questionnaires accurately.
• Followed ABS collection and privacy procedures on every visit to protect data quality and confidentiality at the source.
• Used ABS mobile applications to receive workload, track completion and submit daily administrative reporting.

Hyderabad, India

Owned daily ingestion pipelines for global fund ratings data at Principal Global Asset Management, and found the root cause of a months-old production failure.

• Found the cause of a months-old Morningstar ingestion failure in non-prod. An unpaginated S3 listing stopped at 1,000 objects, so only the largest-backlog environment failed. A one-line fix restored the pipeline and unblocked the release.
• Built AWS infrastructure as code with CDK in Python across S3, Lambda and Glue, with assertion tests for runtime versions, IAM permissions and encryption.
• Owned daily ingestion of Morningstar and Irish-domiciled fund ratings into product tables used by client-facing teams across equities, fixed income and real estate.
• Maintained a three-stage Lambda pipeline for validation, transformation and post-processing checks, with file and pipeline state tracked in DynamoDB for safe reruns.
• Rewrote Informatica mappings and SQL workflows and changed how the jobs read from S3, which cut ETL runtimes.
• Ran internal sessions on AWS CDK and Python testing and received a team award for most active contributor.