Leo Dong

Leading applied ML and data science across multiple brands and 5 markets (AU, UK, ES, DE, JP).

• Own CRM decisioning: who we contact, when, and with what.
• Build churn models feeding retention KPIs, and upsell models driving monetisation.
• Define the experimentation framework and ML decisioning approach.
• Set the data science roadmap and deliver it end to end, using agentic AI to run multiple workstreams in parallel.

First data scientist covering end-to-end funds flow at Checkout.com, a global payments processor.

• Defined north star metrics and net revenue targets, and shaped cost optimisation initiatives, for the Financial Experiences pillar — a cross-functional team of 50+ engineers and PMs overseeing all funds flow through the platform.
• Enabled the launch of a new product distributing over US$100m daily by modelling liquidity, the cost base across global jurisdictions, and future revenue.
• Led rapid-response analytics and controls for EU/UK/MENA/US regulatory compliance, supporting global expansion and new operating licences.

• Built a reusable agentic LLM engine in tested Python (pytest, CI) that generates and cross-links a 40,000-record historical corpus for a Flutter iOS timeline app (public TestFlight beta).
• Designed its LLM evaluation framework (schema checks, tiered LLM-as-judge, significance-tested prompt A/B tests), holding wrong-subject records to 0.25% on a 7,900-record batch.

Built customer models and MLOps pipelines for one of the UK’s largest supermarket and retail groups.

• Increased Google Ads return on advertising spend (ROAS) by 200%+ and conversion value by 150% using propensity models to target high-potential-value customers.
• Designed and implemented a Docker-based ML scoring pipeline that brought model scoring in-house, reducing third-party costs and increasing internal model adoption.
• Collaborated with analysts and SMEs to create a first-of-its-kind, holistic, customer-level view of promotional spending, now used across all marketing to inform strategic decisions.

Led data science strategy across a portfolio of fintech start-ups.

• Led data science strategy for Vonto, a B2B small business analytics start-up, directing the product build and running A/B tests that lifted conversion by 11%.
• Consulted on personalisation methodology for the initial build of its core recommendation system as part of the founding team of Cheddar, an AI-powered deal discovery app co-developed with BCG Digital Ventures.

Supervised industry projects and coursework for the Master of Data Science and Innovation programme.

• Guided master's students through delivering industry-sponsored projects, such as water chemical monitoring models for Sydney Water, and marked submissions.

Delivered projects within a data science team, using statistics and machine learning to build (end to end) and consult on bespoke, customer focused products at Australia's largest financial institution.

• Co-built and deployed Bill Sense, a customer-facing app feature predicting bill amount, frequency and recurrence for over 7 million customers, and notifying them when their bill pricing changed.
• Defined the eligibility matching that connected customers to government rebates, unlocking A$500m+ for customers; the team won the 2019 Gartner Eye on Innovation Award.
• Built economic forecasts from longitudinal customer data for the Australian Federal Cabinet during the COVID-19 pandemic.

Provided analytics for the strategic sourcing and performance management centres of excellence.

• Analysed trends in technology usage and contingent employee costs.
• Supported master services agreement negotiations with enterprise IT providers.

Managed evidence and submissions for the Royal Commission into Institutional Responses to Child Sexual Abuse.