I turn ambiguous operational problems into dependable technical outcomes.
At Auckland Transport, I work across public transport data, scheduling platforms, vendor delivery, testing, and release change. I am trusted to understand how changes move through connected systems, surface risk early, and coordinate the work needed to publish accurate customer information.
Alongside that role, I build practical products and automation that solve problems I understand first-hand. My next step is formal ownership across a platform, integration, solution, or technical implementation outcome - building on work I already do without overstating the title I hold today.
Own and coordinate the end-to-end change process for monthly timetable and network releases, from small metadata corrections to changes affecting large parts of Auckland's public transport network.
02
Systems and data judgement
Work as a subject matter expert across GTFS schedule data and the IVU scheduling platform, validating changes and considering downstream use in AT Mobile, Google Maps, operational tools, and the enterprise data warehouse.
03
Builder behaviour
Identify repeatable operational problems, shape practical solutions, and build working tools - including automated GTFS preparation workflows and Reconcile, a deployed feed-comparison product.
From analysis into ownership
Experience
Business Analyst, Public Transport Data & Customer Information
Auckland Transport
May 2022 - Present
A technical and operational role centred on getting complex schedule data and network change safely through connected platforms and into the customer-facing information people rely on.
Release ownership: own and coordinate the end-to-end change process for monthly timetable and network releases through Azure DevOps, aligning requirements, dependencies, testing, vendor activity, and release evidence.
Downstream correctness: assess and validate GTFS schedule changes with SQL, Excel, and Python, considering their effect across AT Mobile, Google Maps, operational tools, and the enterprise data warehouse.
Technical translation: act as the primary liaison with IVU Traffic Technologies for configuration, upgrades, issue resolution, and service performance, turning operational needs and observed behaviour into actionable technical work.
Procurement and assessment: contribute to RFI and RFP processes for product procurement and vendor assessment, helping define requirements, evaluate responses, and document findings for decision-makers.
Risk and quality: clarify expected system behaviour, develop test cases, coordinate UAT, and resolve data or software-dependent issues before they become customer-facing problems.
Workflow redesign: built AI-assisted Python and SQL workflows that load GTFS data into SQL and generate tester-ready outputs, then extended them to ingest feeds directly from web links.
Future capability: conduct service reviews and process mapping, support business cases, evaluate capability options, and contribute SME judgement to initiatives including National Ticketing, City Rail Link, and Community Connect / OpenLoop.
Integration exposure: work with external developers on data-publishing services and integrations, bringing schedule-data context to solution and delivery decisions.
Analyst
Retail Performance
September 2016 - November 2021
Converted mystery-shopping and market-research data into client-ready findings, recommendations, executive summaries, and visual reporting for non-technical audiences.
Designed and maintained web-based forms and data-collection structures to improve consistency, reporting quality, and downstream analysis.
Worked directly with remote clients to clarify problems, explain evidence, and turn analysis into practical action.
Web Administrator
Digital Works
January 2016 - September 2016
Produced customer-focused, SEO-aware content and supported web administration, analytics, and content maintenance for client sites.
Selected work
Case studies
These projects show how I frame problems, make technical and product decisions, and stay accountable for whether the result works.
A browser-based workbench that helps reviewers distinguish meaningful timetable and network changes from the identifier and feed-window churn that makes raw file comparison misleading.
Problem
Two valid GTFS releases can look radically different even when the service passengers experience has barely changed. A naïve file diff creates noise rather than useful review evidence.
Why it mattered
Reviewers need to focus limited attention on real changes and understand why the tool believes two entities match before accepting a result.
My role
Identified the problem, defined the product, modelled the GTFS domain, made the product and technical trade-offs, and independently built, debugged, iterated, and deployed the working application using AI-assisted development.
Key decisions
Compare service active on a selected date, not just rows in two files.
Use indexed, confidence-scored matching across routes, stops, trip patterns, scheduled trips, and shapes.
Show match evidence and before-and-after detail so conclusions remain inspectable.
Solution boundary and data flow
Meaningful comparison without uploading operational data
Parsing, matching, and classification stay inside the browser. A Web Worker separates the comparison engine from the interface, while searchable views and portable exports turn engine decisions into review evidence.
Baseline + candidateGTFS ZIP feeds
→
Local browser boundary
Parse + validateCalendar expansion and compact indexes
→
Match + classifyDeterministic worker engine
→
Review evidenceViews, maps, JSON, CSV, HTML
Trade-offs
Keeping data local improves privacy and removes a backend dependency. The trade-off is no server-side history; results must therefore be clear, self-contained, and exportable.
Validation and outcome
Large-feed processing, paginated results, match confidence, evidence views, and exports were designed as quality controls around the core comparison. The outcome is a deployed, coherent tool that turns a domain-specific review problem into an inspectable workflow.
What this demonstrates: product definition, systems thinking, domain modelling, explicit trade-offs, edge-case reasoning, and ownership through to a working outcome - not just the ability to write application code.
Selected interface views. Select any screenshot to open the full-size image.
Input boundary: choose baseline and candidate feeds for a controlled comparison.Domain view: review route calendars, stopping patterns, and active trips.Explainability: inspect classification, match evidence, and before-and-after data.Spatial evidence: compare previous and current route geography.
A responsive digital gallery and web edition of an illustrated memoir, designed and shipped for New Zealand artist Blake Twigden.
Problem and role
Bring artwork, biography, current notices, Whitford Bird Garden history, and private enquiries into one coherent experience without letting the interface compete with the work. I shaped the site, content structure, interactions, and delivery end to end.
Key decisions
Used reusable content structures and an artwork-led visual hierarchy; designed keyboard-accessible gallery interactions and reduced-motion support; separated the secure enquiry workflow from browser code.
Quality and constraints
Balanced high-resolution visual material with responsive image delivery, page performance, SEO foundations, accessible full-screen viewing, and light/dark themes.
Outcome
Shipped the official six-page site, including structured metadata, canonical URLs, sitemap and robots directives, Cloudflare Turnstile, and a secure contact workflow.
What this demonstrates: independent product delivery, translating a loosely defined need into a usable system, and taking responsibility for content, interaction, accessibility, security, and release quality together.
Selected interface views. Select any screenshot to open the full-size image.
Artwork-led home page introducing Blake and his practice.Chapter-based web edition of Blake's illustrated memoir.Responsive selected-artworks gallery with accessible full-screen viewing.Current Works feature for the latest collection.
Earlier evidence
Analysis, data design, and communication
Netflix usage analysis
Explored a shared account's 2016-2020 usage in Python, then turned the findings into a written Power BI report and interactive visualisations.
Designed client-facing reporting and data-collection structures that turned operational observations into consistent evidence and concise recommendations.
Built a source-grounded assistant from ITIL training material, notes, and sample exams to generate practice questions, assess answers, and point back to supporting material.
A bounded example of using generative AI to improve a real workflow while preserving human review and source traceability.
Open the interactive Power BI example
Practical capability
Capabilities
Solution & systems
Systems analysis and system behaviour
Data flows, dependencies, and downstream impacts
Technical requirements and solution options
Workflow redesign and integration support
Data quality, reliability, and delivery risk
Product & delivery
Problem framing and discovery
Prioritisation and acceptance criteria
Test cases, UAT, and release validation
Change control and end-to-end coordination
Product thinking and independent delivery
Technical
SQL, Python, Excel, and Power BI
GTFS and IVU.plan
Azure DevOps and Visio
TypeScript and React (project experience)
Microsoft 365 Copilot, ChatGPT, Claude, Gemini, and Codex
AI-assisted development and grounded assistants
Business & implementation
Cross-functional facilitation
Vendor liaison and issue resolution
RFI/RFP support, product procurement, and vendor assessment
Operational analysis and service review
Business cases and capability evaluation
Client-facing analysis and communication
Foundation and development
Education
Bachelor of Computer and Information Sciences
Auckland University of Technology · 2015
Double Major - Software Development and IT Service Science
Professional development
ITIL Foundation - PeopleCert, 2026 Project Management Essentials (PMBOK) - Lumify, 2024 Data Analyst in Python / SQL - Dataquest, 2020
Ongoing self-directed development in AI, LLM workflows, automation, and modern digital delivery.
