About Projects Contact GitHub
Portfolio

Building things
that solve real problems.

Developer and automation enthusiast building platforms, cross-platform apps, and AI-powered tools that make complex workflows feel effortless. Based in Adelaide.

15 Years in FM
End to End Build, Ship & Scale
AI Native Agents & Automation
01

About

I build platforms and tools that bridge the gap between complex systems and the people who use them. With over 15 years in enterprise operations — spanning customer-facing roles, facilities coordination and management, scheduling, business systems, project management, and data & analytics — I've learned to spot the problems before they become expensive.

My projects span full-stack web platforms, cross-platform mobile and desktop apps, AI agent orchestration, data pipelines, and NLP classification engines. I'm drawn to problems where the right tool — built properly from the start — can eliminate hours of manual work. I'm focused on building the right thing for the problem, whether that's a fully architected platform built to last or a quick script hacked together at midnight to get the job done. When I'm not coding, I'm probably at the gym (where I've automated my booking and workout tracking), or keeping tabs on how much my dog barks while I'm out — it's more than I'd like thanks to the neighbourhood terrorist (Shadow), but he's still a certified good boy (eddieisagoodboy.com).

Languages

  • Python
  • JavaScript / HTML / CSS
  • T-SQL / DAX / M
  • Shell / Bash

Frameworks & Platforms

  • Flask / FastAPI
  • React Native / Electron
  • HTMX / Alpine.js / Tailwind
  • Prefect / Data Pipelines
  • Stripe / Resend

AI / Data

  • Claude API / Claude Code
  • NLP / Keyword Classification
  • Computer Vision / LLM Analysis
  • Pandas / SQLAlchemy
  • PostgreSQL / MSSQL / SQLite
  • Chart.js / Plotly

Infrastructure

  • Docker / DigitalOcean
  • Git / GitHub
  • Azure / AWS
  • Gunicorn / Caddy
  • REST API / Swagger
  • GitHub Actions / CI/CD
02

Projects

LK
Client Project 2026

LK Design & Build

Complete digital presence for an Adelaide design & construction studio. Six interactive website prototypes exploring different design languages — from scroll-driven video showcases with canvas frame extraction to animated infographic overlays. Includes a full branding kit with five logo concepts, business card mockups, vehicle wraps, signage, and brand guidelines.

HTML / CSS / JS Canvas API Scroll Animations SVG Infographics AI Image Gen GitHub Pages
View Project
SH
Client Project 2026

Saturn Hire

Six website prototypes for Adelaide's portable facilities hire company. Scroll-driven video reveals with canvas frame extraction, AI-generated animation keyframes, cosmic branding themes, and real business content. From bold construction energy to knowingly absurd luxury — premium design language applied dead-straight to portable toilets.

HTML / CSS / JS Scroll-Driven Video Canvas Frame Engine AI Image Gen GitHub Pages
View Prototypes
IoT / ML 2025

Bark Up

AI-powered dog bark tracking using YAMNet ML classification and Nest Cam integration. Captures forensic evidence (audio/video clips with confidence scoring), logs episodes to a Notion database, and sends Telegram alerts with false positive filtering. Includes iOS Shortcut and Siri integration for on-demand reports.

Python YAMNet ML Google Cloud Notion API Telegram Docker
View on GitHub
Eddie
Live Data Dashboard 2025

Eddie Is A Good Boy

Public-facing live dashboard for the Bark Up ML system. Displays real-time bark episode tracking, daily and all-time statistics, hourly activity charts, 14-day trend analysis, and a dynamic angel/devil mood indicator driven by behaviour data. Streams from the YAMNet classification pipeline to a Flask web frontend.

Flask Live Data YAMNet ML Chart Visualisation Behavioural Analytics
Visit eddieisagoodboy.com
Enterprise Platform 2026 — Present

Triage Classification Engine

Enterprise work order triage platform with a three-path classification engine (equipment lookup, 860+ weighted keyword patterns across 19 disciplines, and LLM fallback), context-aware smart questions that adapt by facility type, and an active learning system that turns agent overrides into rule improvements. Includes a fully configurable form builder with conditional visibility and field inheritance, a declarative output builder supporting 11 payload types with expression-based field resolution, and a PWA mobile-first customer portal with barcode scanning and live order tracking. Handles healthcare, education, justice, and housing environments across multiple contracts.

Python / Flask AI / LLM (Multi-Provider) Classification Engine Form Builder Output Builder PWA SQLite / Databricks Docker Azure CI/CD
Real-time Dashboard 2024 — Present

Operations Dashboard

Wall-mounted rotating display that federates data from multiple sources — SQL Server, AWS Athena, Jira, and Power BI semantic models. Auto-rotates between KPI pages with configurable refresh intervals and caching. Features encrypted credential management, custom DAX queries, and can be bundled as a standalone executable for deployment on dedicated display hardware.

Python / Flask Chart.js SQL Server AWS Athena Power BI API PyInstaller
Automation 2025

Workout Intelligence

No accessible public-facing API? No problem — that just depends on what you consider "accessible" and "public-facing." Point mitmproxy at the app, watch the calls roll by, and suddenly you've got yourself a perfectly accessible, public-facing API. Reverse-engineered endpoints feed a nightly GitHub Actions pipeline that pulls gym programming, generates hypertrophy accessory suggestions via Claude Haiku, and delivers personalised workout plans through Telegram and Siri — all for roughly $0.003/week.

Python mitmproxy API Reverse Engineering Anthropic API GitHub Actions Telegram Siri Shortcuts
View on GitHub
Automation 2025

Session Auto-Booker

Built on the same reverse-engineered API groundwork as Workout Intelligence, but this one goes both ways — reading session availability and writing bookings back through endpoints that were never meant to be called outside the app. A Flask web UI lets you pick preferred time slots, then a background scheduler fires every Sunday at 8 PM to lock them in automatically. Encrypted credential storage, Dockerised deployment behind Caddy HTTPS, and confirmations pushed via Telegram and Siri.

Python / Flask API Reverse Engineering Scheduler Encryption Docker Caddy HTTPS Telegram
Enterprise Tool 2025

Auto Scheduler

Visual flow builder for creating and deploying automated data pipelines. Build flows from Python functions, SQL procedures, or generate full Prefect syntax — then deploy directly to Prefect Cloud. Includes a persistent agent service running on the VM that executes and monitors scheduled flows.

Python Prefect T-SQL Flow Builder UI VM Agent
Full-Stack App 2025

Footy Tipping

Full-stack AFL tipping prediction app with FastAPI backend, PostgreSQL database, and Docker containerisation. Includes Alembic migration pipeline, SQLAlchemy ORM modelling, and pytest suite. Enterprise-grade architecture for a decidedly non-enterprise pursuit.

Python / FastAPI PostgreSQL SQLAlchemy Docker Alembic
Data Science 2024

Data Usage Analysis

Jupyter-based geographic data usage analysis with Pandas and visualisation libraries. 3,300+ lines of exploratory analysis, statistical breakdowns, and interactive charts mapping consumption patterns by location.

Python Jupyter Pandas Data Visualisation
Productivity Tool 2024

AutoTimer

Passive productivity tracker that monitors where your time actually goes — tracking active browser tabs, applications, email activity, and keyword patterns to map how attention flows across work streams. Built before tools like Rize existed, to answer the question: "Where does my productive time really go?" Surfaces deep work vs shallow work ratios, context-switching frequency, and time-per-project breakdowns without manual timers or start/stop buttons.

Python Browser Monitoring App Tracking Time Analytics Productivity Analytics
View on GitHub
API Integration 2024

Spotify Playlist Generator

Programmatic Spotify playlist creation via the Spotipy API. Jupyter notebook that builds curated playlists based on configurable criteria — genre, tempo, energy, and mood parameters fed through Spotify's audio features endpoints.

Python Spotipy API Jupyter OAuth 2.0
03

Get in Touch

Interested in working together or want to discuss a project? Reach out.