Data Scientist
2020 — PresentCommercial Analytics, BHP
- Built a real-time vessel scheduling system that increased iron ore sales by 18M tonnes annually ($900M value) through optimised fleet allocation and inventory management.
- Shipped end-to-end LLM workflows automating complex assay data extraction and contract reconciliation — eliminated 600+ hours of manual effort annually and recovered $2.5M in revenue.
- Deployed LLM models on AWS SageMaker that classified $2.7B of off-contract spend, reducing procurement compliance risk by 35%.
- Led Snowflake platform optimisation for 200+ Commercial users — 50% reduction in compute costs alongside data governance + automated CI/CD pipelines.
- Supported migration from IBM CP4D to AWS, implementing enterprise security frameworks that enabled 15+ analytics workstreams and reduced infrastructure cost by 40%.