Projects

Selected projects demonstrating impact across analytics, engineering, operations, and product development. View full Project Portfolio with craft competency mapping.

Analyticado

Centralised analytics hub with ~500 daily active users

Internal website integrating Looker Enterprise metrics and external analytics outputs into a single discovery platform. Features a custom search engine, Looker API integration, embedded dashboards, and UX-focused design.

Impact

  • ~500 daily active users
  • 10x increase in analytics engagement
  • One-stop search for all analytics resources
Next.jsReactLooker APITypeScriptBigQuery

Python Framework

Shared analytics automation platform with ~10 contributors

Reusable Python framework automating data pipelines across Looker, Slack, BigQuery, and email. Includes ML utilities, GCP Cloud Functions, CI/CD via GitLab, and secure credential management.

Impact

  • ~20,000 lines of code
  • ~10 contributors across the team
  • ~5 GCP Cloud Functions in production
PythonGCPBigQuerySlack APIGitLab CI/CD

Cost to Serve Model

$1M annual savings through operational optimisation

Analytical model identifying cost inefficiencies in the client's fulfilment operations. Delivered actionable recommendations that led to significant annual savings and a client-facing self-service version.

Impact

  • $1M annual savings for the client
  • Client self-service capability delivered
  • Model adopted across operations leadership
SQLBigQueryLookerGoogle Sheets

Water Limit Analysis

~$550K saved with ~$2M+ further potential

Geo-cubing analysis of water distribution limits across the client's fulfilment network. Identified optimisation opportunities in capping and routing that were adopted by operations teams.

Impact

  • ~$550K/year in direct savings
  • ~$300K-~$2M/year in recommended optimisations
  • Adopted by operations leadership
SQLPythonBigQueryStatistical Analysis

Routes Predictor

ML-powered route forecasting for labour planning

ARIMA-based forecasting model predicting delivery routes to support labour planning. Integrated with GCP Cloud Functions for automated execution and AppScript for stakeholder access.

Impact

  • More accurate interim predictions than existing heuristics
  • Automated forecasting via GCP pipeline
  • Guided data-driven labour planning for the client
Pythonscikit-learnARIMAGCP Cloud Functions

OSP Insights Platform

Partner reporting platform with ~3,000 monthly report runs

Redesigned landing pages and navigation for the partner-facing Looker reporting platform. Structured visual flows to guide users from summary to detail-level insights across multiple partners and 100+ sites.

Impact

  • 2x month-over-month usage increase
  • ~3,000 report runs per month on platform
  • Produced the most used and impactful reports; top report ~2,000 runs per month (In the Bag)
LookMLLookerSQLBigQuery

Capture - Mobile Application

Cross-platform photo sharing app launched on iOS & Android

Independently designed, developed, and launched a consumer mobile application focused on delayed-reveal photo sharing. Full lifecycle ownership from concept and branding through to monetisation and App Store Optimisation.

Impact

  • Successfully launched on both app stores (January 2026)
  • ~100 downloads with ongoing growth
  • Full monetisation: subscriptions, one-time purchases, and ads
FlutterDartSupabasePostgreSQLFirebaseFigma
View website

K-AI - Conversational Analytics Assistant

AI-powered analytics discovery across dashboards, docs, and Slack

Internal AI assistant enabling stakeholders to discover and understand data, dashboards, and documentation using natural language. Reduces friction in analytics discovery and enables non-technical users to query complex data ecosystems. Built with Vertex AI and hybrid search.

Impact

  • Reduced friction in accessing analytics and knowledge
  • Enabled non-technical users to query complex data ecosystems
  • Centralised fragmented knowledge (Looker, GCS docs, Slack) into one interface
Vertex AIBigQueryGCPLooker API

Workload Forecasting ML Models

Network-wide staffing optimisation across >100 sites

Machine learning models for predicting warehouse workload and staffing needs across the global fulfilment network. Enabled proactive resource allocation and labour management.

Impact

  • Network-wide staffing optimisation
  • Deployed across 100+ sites (CFCs, MFCs, in-store fulfilment)
  • Serving ~10 partner organisations
Pythonscikit-learnBigQueryGCP