Corporate highlight 1
Enterprise AI & Machine Learning Delivery
- Challenge
- Identify potential shadow-economy activity and suspected undeclared income across millions of online data sources, enabling the ATO to target limited investigative resources towards the highest-risk entities.
- Approach
- Led a multidisciplinary team of 6, comprising 3 data scientists and 3 business specialists. Established a secure cloud analytics capability and developed Linux Bash and Python web-crawling, indexing and machine-learning pipelines to analyse 2 million URLs. Built predictive risk models and a decision-optimisation layer to rank entities and prioritise investigative actions.
- Outcome
- Identified approximately 1,900 entities for investigation, enabling granular risk-based targeting, more effective allocation of enforcement resources and referrals to specialist task forces. Delivered the three-year programme in 2.5 years, completing it six months ahead of schedule, and subsequently applied the remaining contract period to optimise other data science capabilities.
- Related
