Nearly two decades across Sydney and Perth: incubating 100+ startups, growing a cabin business from idea to profit and now shipping with AI.
I started my first business in France straight out of high school. In my twenties I moved to Sydney and joined Pollenizer, Australia’s first web incubator, where I became Head of Product and built the team that launched 100+ startups. I coached 30+ of them directly, using a program we developed and later published as the book Startup Focus. It worked: one of those startups, Spreets, sold to Yahoo!7 for AU$40M.
In 2019 I moved to Perth and started Heyscape inside RAC’s innovation hub, growing it from an idea into a profitable business with 25 off-grid cabins across seven sites.
Today I build my own AI products and lead AI projects on contract. Most recently, at Respondent.io, I built an AI platform that runs their recruitment campaigns.
Weekly gatherings of things I read on the state of the industry.
Anthropic’s S-1 gives us real numbers, the market hunts for cracks in the AI build-out, North Korea takes $6.3bn from Binance, and two labs claim to disprove a 25-year-old hypothesis.
Demystifying buzz words, and other rants
Four products I designed, built and run on my own. Each case study covers the architecture, the numbers and what I learned.
Made AI coding usage measurable across every tool.
A Mac menu bar app and web dashboard that track tokens, cost and output across Claude Code, Codex, Cursor and Mistral Vibe.
Production grade setup in one minute.
A free Mac app that gives each new project its own AWS account, GitHub repo, keyless deploys and agent guardrails.
Made flaky, metered AI video reliable enough to sell.
Turns a handful of photos into a cinematic short film in minutes.
Made live voice AI hold up on a real phone line.
An AI receptionist for Australian tradespeople and real-estate agents: it answered missed calls, triaged every message and called people back.
Open-source things I’ve made, and other people’s projects I keep picking up.
Made herepierresauvignon /house-rules
Nine house rules for AI coding agents: sound human, aim high, tell the truth, keep it cheap, delegate well, cut friction, respect the code, stay secure.
Markdown GitHub ↗ MITpierresauvignon /Property-map
A shared notebook for two people buying a house together: every place you’re considering on one map, with both your scores, comments, photos and visits.
JavaScript GitHub ↗A few favourites from my photography portfolio on 500px.
My last three posts on X.
Read on X ↗That kinda of visualisation UI reminds me of the Winamp’s skins (skins.webamp.org). Completely useless but somewhat desirable :D
— Pierre (@pierre_s) October 8, 2026
Read on X ↗Shit is about to hit the fan.
— Pierre (@pierre_s) September 17, 2026
Read on X ↗@Scobleizer I don’t understand how/why this👇 is not a thing in VR yet? Seems to me like an obvious banger for Apple (and others) yet no signs of it.
— Pierre (@pierre_s) September 3, 2026
Based in Perth, happy to work remote or on-site.
pierre.sauvignon at gmail dot com