KC

Kevin Chang Senior Data Scientist · Banking Pricing · AI · Auckland, New Zealand

Turning pricing, AI, and customer intelligence into business decisions.

Kevin Chang helps financial services leaders connect advanced modelling with revenue, customer outcomes, conduct risk, and clear executive action.

Explore experience Resume LinkedIn
Kevin Chang, Senior Data Scientist
Now · Senior Data Scientist, ASB Bank

12+

years

Applied analytics across banking, policy, consulting, and research

ASB Bank
Retail banking pricing, customer intelligence, and model assurance
Ph.D.
Statistics depth with public-policy and commercial delivery

Data science leader delivering measurable business impact

12+ years driving revenue growth, customer value, and strategic decision-making through advanced analytics, machine learning, and data products. Currently at ASB Bank optimising pricing strategy and driving customer intelligence initiatives. Rare combination of commercial expertise, technical depth, and policy modelling knowledge—proven ability to translate complex models into clear, actionable insights for C-suite stakeholders. Specialist in customer lifetime value systems, predictive analytics, generative AI applications, and large-scale microsimulation. Recognised mentor building high-performing analytics teams. Experience spans financial services, government policy, and consulting.

Experience
12+ years turning models into decisions Financial services, government policy, consulting, and research delivery with a consistent focus on usable decision support.
Banking focus
Pricing work connected to P&L, customer outcomes, and risk Current ASB focus spans pricing models, product performance, conduct considerations, and post-launch optimisation.
Customer intelligence
Customer growth and retention modelling Customer lifetime value, segmentation, churn prediction, and recommendation systems translated into business planning.
AI capability
ML and GenAI with practical guardrails Predictive analytics plus LLM analysis of open-ended survey responses, customer sentiment, and signal extraction.
Risk discipline
Model assurance background Validation experience across risk, credit, and operational models with regulatory expectations.
Leadership
Executive-ready communication and mentoring Turns technical complexity into clear recommendations for senior stakeholders while building analytics capability in teams.

Experience

From statistical consulting and government policy modelling to banking pricing and model assurance at ASB.

Jan 2026 – Present

ASB Bank

Auckland, New Zealand

Senior Data Scientist

Lead pricing strategy by integrating market intelligence, competitor dynamics, and advanced analytics to drive revenue growth and competitive positioning.

  • Drive continuous improvement of pricing strategies, models, and processes to enhance customer experience and business performance.
  • Develop and implement pricing models and analytical frameworks to support new product features or enhancements aligned with customer and market needs.
  • Monitor product performance, including pricing impact on revenue, profitability, and customer outcomes, and provide actionable insights to optimise P&L performance.
Show 3 more
  • Apply simulation-based analytical frameworks—drawing on a background in large-scale policy microsimulation—to model the distributional impact of pricing changes across customer segments before deployment.
  • Evaluate post-launch performance using advanced analytics, recommending adjustments to pricing, product features, and customer strategies to improve outcomes.
  • Identify, assess, and proactively manage risks associated with pricing decisions, including regulatory, conduct, and model risks.

Pricing · Analytics · Revenue Optimisation · Advanced Statistical Modelling

Jan 2023 – Jan 2026

ASB Bank

Auckland, New Zealand

Data Scientist

Lead strategic analytics initiatives, drive customer growth, and business innovation across the bank.

  • Customer Intelligence & Growth: Developed and deployed predictive models for customer lifetime value, segmentation, churn prediction, and product recommendation, delivering actionable insights that inform strategic planning and drive measurable revenue growth.
  • Generative AI: Applied large language models (LLMs) to analyse open-ended survey responses, identifying common themes and customer sentiments driving NPS results and customer experience insights.
  • Data Product Leadership: Partner with product, design, and engineering teams to build scalable insights platforms and analytics tools that enable data-driven decision-making across business units.

Customer Intelligence · LLM · ML · Data Products

Jan 2021 – Jan 2023

ASB Bank

Auckland, New Zealand

Model Assurance Specialist

Provided independent, rigorous validation of critical bank-wide models to ensure regulatory compliance and risk management.

  • Conducted comprehensive end-to-end validations of risk, credit, and operational models, assessing methodology, data quality, implementation, and ongoing performance monitoring.
  • Challenged model assumptions and methodologies through rigorous statistical analysis, identifying limitations and recommending enhancements that strengthened model reliability.
  • Collaborated with model developers, risk managers, and senior stakeholders to ensure models met regulatory standards (RBNZ/APRA requirements) and internal risk policies.
Show 1 more
  • Delivered clear, comprehensive validation reports to senior management and governance committees, facilitating informed decision-making on model approval and risk mitigation.

Model Validation · Risk · Regulatory Compliance · Advanced Statistical Modelling

Jan 2019 – Jan 2021

The New Zealand Treasury

Wellington, New Zealand

Modelling Analyst

Developed sophisticated analytical tools supporting government policy development and fiscal planning.

  • Policy Impact Modelling: Built and maintained microsimulation models to analyse tax and welfare policy impacts, directly informing Budget decisions and Ministerial recommendations.
  • Data Integration & Engineering: Integrated complex survey and administrative datasets (IDI, HES) to create robust analytical foundations for policy evaluation.
  • Product Development: Created interactive R Shiny dashboards and internal R packages that streamlined policy analysis workflows and improved accessibility of insights for policy analysts.
Show 1 more
  • Stakeholder Collaboration: Worked closely with policy teams, Statistics NZ, and other government agencies to ensure analytical outputs aligned with strategic policy objectives.

Microsimulation · Policy · R Shiny · Government

Jan 2014 – Jan 2019

Statistical Consulting Centre, University of Auckland

Auckland, New Zealand

Statistical Consultant

Delivered end-to-end analytical consulting solutions across academia, government, and industry, building custom solutions for complex statistical challenges.

  • Provided statistical expertise to 60+ clients across academic, government, and commercial sectors, translating business questions into rigorous analytical approaches.
  • Designed and deployed web-based analytical tools and custom R packages for clients, including government agencies and research institutions.
  • Conducted advanced analyses spanning experimental design, survey methodology, longitudinal modelling, and causal inference.
Show 1 more
  • Led R programming workshops and training sessions, building analytical capability among researchers and practitioners.

Consulting · R · Statistics · Training

Selected work

End-to-end delivery: problem framing, modelling, and a usable tool at the end of it.

2026 · Policy model

Simulation Modelling for A Better Start

Simulation model evaluating how early-life interventions in literacy, growth, and wellbeing influence long-term and equitable outcomes. Models life-course impacts and delivers results through an interactive Shiny web app.

R · Shiny · Microsimulation · MoE

Problem
Policymakers needed to see how early-life interventions in literacy, growth, and wellbeing shape long-term and equitable outcomes.
Approach
Built a life-course microsimulation model and delivered it to non-technical users through an interactive Shiny web app.
Impact
Lets MoE / COMPASS stakeholders test intervention scenarios and equity impacts before committing policy.

2026 · AI system

Multi-Asset LLM Trading Bot

Python · LightGBM · GARCH · HMM

Problem
Agentic decision systems need clear controls, observability, and failure boundaries before they can be trusted with live actions.
Approach
Built a 3-stage LLM pipeline over a quant stack (LightGBM, GARCH volatility, HMM regime detection), with mechanical exits, circuit breakers, and monitoring.
Impact
944-test suite, CI/CD auto-deploy to a self-hosted NAS, read-only dashboard, and layered risk gates for BTC / ETH / SOL spot execution.

2026 · AI system

Personalised Daily Briefing Bot

Python · Docker · OpenRouter · SQLite

Problem
Staying on top of NZ news, markets, and bank rate moves means scanning many sources every day.
Approach
Self-scheduling Telegram bot with an evolving memory store and IQR-based anomaly detection across currencies, indices, and bank mortgage / term-deposit rates.
Impact
Three personalised briefings a day at ~$0.03/day running cost; runs in Docker on a NAS with no external scheduler.

2021 · Decision app

Automated Psychometrics

R · Shiny · Rasch Analysis · Psychometrics

Problem
Reproducible psychometric analysis (Rasch, DIF, equating) required specialist coding, putting it out of reach for most assessment teams.
Approach
Built an open-source R Shiny app exposing the full reproducible workflow through a no-code interface; peer-reviewed and published in PLoS ONE.
Impact
Adopted by 60+ educational institutions worldwide; 1,200+ citations.

More public work

  • 2021

    Interactive Highcharter visualisation of pandemic trends, used by NZ media and health communications to inform public understanding of outbreak dynamics.

  • 2021

    Income Explorer

    Policy model

    The Income Explorer is an interactive app that models the relationship between wages and disposable income, assessing the impact of the New Zealand tax and welfare system on family incomes and work incentive indicators like effective marginal tax rates.

  • 2019

    Vulnerable Children Investment Approach

    Policy model

    Integrated IDI administrative data to identify causal pathways from family violence to child outcomes. Findings informed $50M+ policy investment decisions on prevention and support interventions.

  • 2018

    Microsimulation platform analysing 50+ policy interventions to identify high-impact factors for child wellbeing. Interactive tool used by 20+ NZ government policy teams.

  • 2018

    Network analysis and visualisation of Pacific region development aid flows, informing NZ foreign aid strategy and regional partnership priorities.

  • 2017

    Census-based microsimulation enabling policy teams to test 100+ socioeconomic scenarios. Reduced policy evaluation cycle time from months to days.

Skills

Grouped by capability. Deeper expertise first; supporting tools beneath.

AI / ML / GenAI

  • Machine learning
  • Predictive analytics
  • Generative AI / LLMs

LLM survey & sentiment analysis · Churn prediction · Customer segmentation · Product recommendation

Pricing / Banking / Optimisation

  • Pricing strategy
  • Revenue optimisation
  • Customer lifetime value

Customer intelligence · Model assurance & validation · Regulatory model risk

Data Engineering / Analytics Platforms

  • SQL
  • Snowflake
  • Databricks

dbt · Shiny dashboards · Data products · Git

Programming / Modelling

  • R
  • Python
  • Advanced statistical modelling

Microsimulation · Experimental design · Data visualisation

Communication / Stakeholder Impact

  • Executive insight translation
  • Stakeholder communication
  • Mentoring

Research · Consulting

Research

Peer-reviewed statistics, modelling and open-source tools. The technical depth behind the commercial work.

citations
1,200+ citations autopsych, PLoS ONE (2021)
institutions
60+ institutions using the open-source autopsych Rasch analysis tool
  1. 2021

    Courtney, M. G. R., Chang, K., Mei, E., Meissel, K., Rowe, L., & Issayeva, L. (2021). autopsych: An R Shiny Tool for the Reproducible Rasch Analysis, Differential Item Functioning, Equating, and Examination of Group Effects. PLoS ONE. Open-source tool adopted by 60+ educational institutions globally; 1200+ citations.

  2. 2019

    Shackleton, N., Chang, K., Lay-yee, R., D'Souza, S., Davis, P., & Milne, B. (2019). Microsimulation model of child and adolescent overweight: making use of what we already know. International Journal of Obesity. Policy modelling supporting NZ child health intervention strategies.

  3. 2019

    Zhao, J., Mackay, L., Chang, K., Mavoa, S., Stewart, T., Ikeda, E., ... & Smith, M. (2019). Visualising combined time use patterns of children's activities and their association with weight status and neighborhood context. International Journal of Environmental Research & Public Health.

  4. 2019

    Sutherland, K., Clatworthy, M., Chang, K., Rahardja, R., & Young, S. W. (2019). Risk factors for revision anterior cruciate ligament reconstruction and frequency with which patients change surgeons. Orthopaedic Journal of Sports Medicine.

  5. 2019

    Mackenzie, B. W., Chang, K., Zoing, M., Jain, R., Hoggard, M., Biswas, K., Douglas, R. G., & Taylor, M. W. (2019). Longitudinal study of the bacterial and fungal microbiota in the human sinuses reveals seasonal and annual changes in diversity. Scientific Reports.

  6. 2018

    Lay-Yee, R., Milne, B., Shackleton, N., Chang, K. & Davis P. (2018). Preventing youth depression: Simulating the impact of parenting interventions. Advances in Life Course Research. Microsimulation informing evidence-based youth mental health policy.

  7. 2018

    Courtney, M. G. R. & Chang, K. C. (2018). Dealing with non-normality: An introduction and step-by-step guide using R. Test Journal. Practical statistical guidance for 500+ academic practitioners.

  8. 2017

    Chang, K. (2017). Computer generation of designs for two-phase experiments with applications to multiplex experiments in proteomics [Doctoral thesis, The University of Auckland]. Statistical design methodology with applications in high-dimensional biology.

    View record (opens in new tab)

Education

  • 2017

    Ph.D. in Statistics and Biological Sciences University of Auckland

  • 2008

    B.Sc. (Hons) in Bioinformatics University of Auckland

  • 2007

    B.Sc. in Bioinformatics University of Auckland

Languages

  • English · Fluent
  • Mandarin Chinese · Fluent

Outside work

Running, Photography, Travelling