Sydney, New South Wales, Australia
AI is creating winners faster than ever. It’s also creating wreckage faster than anyone wants to talk about. I’ve spent years rolling out AI inside real businesses. That work made two things obvious: most people still aren’t using these tools properly, and a lot of capable people are getting hit from every direction at once. Job loss. Financial shock. Skill drift. Confidence collapse. The ground moving faster than people can respond. So I built TEKVA. An Australian registered charity (PBI, DGR1) that backs capable adults in financial crisis. Emergency relief. AI fluency training. A practical path back to income. We’re AI-native. Frontier models power our triage, case reasoning, financial assessments, and the upskilling we deliver to displaced workers. Small team, real deployment, human oversight on every decision. We back people. Not problems. If you’re building in this space, funding it, or living it, please get in touch. [email protected]
TEKVA is a practical recovery system for working-age Australians under financial pressure. Registered Public Benevolent Institution, DGR1. We help people understand their situation, stabilise the urgent stress, navigate their options, and rebuild. Four pillars: hardship support, financial counselling, AI upskilling, and emergency grants where they’ll change the trajectory. I run operations and build the systems behind it: AI-assisted triage, case infrastructure, vendor-direct payments. AI speeds up the work. Humans make every decision. Why now: economic volatility and AI-driven labour disruption are hitting working-age adults the existing safety net wasn’t built for. We’re building the early-intervention layer that sits between crisis and recovery.
Working at the intersection of copyright, open-source, and LLMs... Where the rules haven’t been written yet. - Mapping how AI models ingest, remix, and repurpose open content at scale - Assessing whether AI-generated work can be licensed, protected, or even owned - Investigating how “open” principles survive (or fail) inside modern AI training pipelines - Advising on emerging standards for AI transparency, attribution, and content rights Basically: trying to define what “open” means in an era where models take everything.
- Built AI copilots that saved teams hours, not just made pretty demos. - Replaced manual workflows with LLM systems people actually adopted. - Benchmarked OpenAI, Claude, Gemini, and OSS models against real-world constraints (not hype). - Set up practical governance so AI was safe and useful. - Turned messy operational problems into clean, automatable flows. - Taught leaders how to think with AI instead of fear it. - Moved companies from “experimenting” to actually running on AI.
Led AI deployment for Zip’s US team - taking AI from idea to implementation. - Successfully rolled out OpenAI Enterprise across key teams - Built and deployed internal GPTs for support, risk, and compliance - Ran PoCs and evaluations on AI models to test real-world performance - Developed AI copilots that actually saved time, not just looked good on paper - Worked with legal and compliance to ensure AI met fintech regulations - Shifted AI from “innovation project” to embedded, everyday ops
Stepped into ABM as an AI Consultant to drag their digital ops into the future. Took their team from clueless to competent with hands-on AI training and custom solutions—because screwing around with half-baked tech wasn’t cutting it. - Ran prompt engineering workshops to stop staff from sucking at ChatGPT and GPT usage. - Built custom GPT applications for CX teams, reducing response times and manual workload - Trained Loras for e-commerce product shots made their online look look hot and sell more - Developed AI tools for sales and research automation, customer insights on steroids. - Whipped up dashboards for sales, freight, and marketplace reporting - data they could actually use. - Used AI to wrestle NetSuite migrations into submission—less chaos, more control.