Consulting Experience

Big 4 AI & Data · 2015–present

A selection of engagements I’m able to discuss publicly. Much of my work, including most of the advanced data-science and machine learning work done in research and reviewer roles, are sensitive; details will not be made available on request.

← Back to profile

Held overall technical ownership of a policy knowledge bot deployed to ~30k users over 200 days, coordinating workstreams across Australia, Hong Kong, the Philippines, and India. Directed architectural and evaluation decisions; managed vendor negotiations and a rolling weekly release cadence.

Ran a global GenAI readiness assessment across ~800k staff, filtering to a few hundred targeted specialists and building capability profiles for each. Developed QA management training for GenAI projects and contributed to the firm’s GenAI market strategy.

Progressed from workstream lead to data science lead over three years on predictive maintenance and mine intelligence for a large WA mining client’s autonomous rail network. Led teams of data engineers and scientists across defect prediction, track dynamics, and planning optimisation. Also worked as a contributing researcher on novel machine learning within the ore-body intelligence team.

Data engineer and data scientist on a skunk-works program to reduce unnecessary field dispatches by predicting from network telemetry whether a scheduled rollout would fix the reported fault. Designed and implemented PySpark pipelines across the client’s maintenance operations and identified and quantified significant systematic waste.

Built backend APIs and data integration pipelines on a Redis Streams architecture, integrating with legacy clinical systems for a cancer care provider’s patient and clinician apps. Worked closely with the client’s head of engineering and coordinated with a team in the UK; additional responsibilities included AWS and Azure automation, information security, and advising on an exploratory analysis of medical treatment bias.

Lead developer embedded within a Victorian government department on an emergency website extension to disburse energy bill credits to qualifying residents. Built on AWS; integrated with four energy distributors, an address dataset provider, and a bill intelligence provider with a custom API. Handled sensitive personal information including passport scans. The site launched over a weekend under TV-driven traffic.

Led a six-month PoC with a large bank’s Threat Intelligence team alongside the local cyber practice and a US team of ex-Los Alamos National Laboratory researchers who brought their patented analytics solution. Successfully demonstrated detection of malicious network activity from global netflow data, underpinning the bank’s decision to build out a dedicated DC hall for real-time monitoring.

Led technical delivery on the ground in San Francisco of an automated validation platform for standardising third-party lab result submissions, requiring deep familiarity with the lab results data domain. Coordinated teams across Melbourne, Poland, San Francisco, and an offshore delivery centre; delivered compliant with US patient data handling requirements.

Designed and built the backend API and data model for a mobile app serving an executive cycling community, ingesting Strava activity data to model power curves for individual cyclists and simulate their performance against each other in famous races.

Delivered orders-of-magnitude performance improvements to an official data integration pipeline for registered training providers supporting medical education, primarily through database query optimisation. The system accepted spreadsheet submissions via a web front-end and was later extended to support regional practice awards.

Inherited, extended, and supported across several collection cycles a data validation and ingest front-end for annual workforce profile collection. Entitlements and summary statistics for every NSW public servant, checked for legislative compliance twice a year, flow through this system.