maclulich.compersonal record / public edit2011—now

a personal timeline of building things

Things I’ve built, broken & learned from.

The tidy version is a list of roles. This is closer to how it felt: launches, late nights, strange bugs, good people, a few sales — and the little moments that made ten years at Audience Republic feel real.

Audience Republic / sampled Slack activityFeb 2016 → Apr 2026
16
17
18
19
20
21
22
23
24
25
26

Not productivity. Just the changing pulse of a company being built.

from the archive

The Audience Republic years come from an AI-assisted reading of ten years of my Slack history. This public edit keeps the character and removes private names, credentials, internal links and moments that belong to other people.

2011—15

goCatch / the prologue

Product + engineering

Software that moved cars through a city

At goCatch, the screen was only the surface. Underneath were live location, dispatch, payments, driver supply and all the sharp edges that appear when a person is waiting on a kerb.

Street-level software. City-scale systems.

Real worldPeople, cars and money moving at once

Hard partMaking live location feel dependable

open that chapter
2015—16

The Medic / built + sold

Founder + builder

A smaller problem, solved all the way

The Medic gave people a direct path from feeling sick to speaking with a GP and receiving a medical certificate. I built it, took it to market and sold it to 1st Group.

Small enough to ship. Real enough to matter.

The productA GP consultation without the waiting room

The outcomeBuilt, launched and acquired by 1st Group

open that chapter
Audience Republic / ten years in the Slack record
2016

Ticketsquad.io / the origin

Co-founder, day one

Two founders and an empty Slack

The first message landed on 11 February. The company was called Ticketsquad.io, there were two people online, an early client to help and the first version of an audience data model to make real.

The whole thing was two people with a big idea.

First message11 February 2016 — the workspace wakes up

First clientA real event and no room for pretend software

First schemaThe audience data model starts taking shape

2017

the rename / first platform

Builder-in-chief

Ticketsquad became Audience Republic

The new name went live with the customary launch-day DNS drama. I was still writing the Clojure, managing design and front-end contractors, and finding out exactly how much Christmas Day a founder can reasonably take off.

It’s quite possible I screwed something up while I was hacking on it in the bush.

Milestoneaudiencerepublic.com goes live

Classic JasonAt least I took most of Christmas Day off

The thesisAudience intelligence from every useful signal

2018

process / integrations / automation

Hands-on builder

The product grew bones

QA, sprints, payment flows, messaging integrations, security processes and the first internal automations arrived. The scrappy thing was becoming a platform — and a company that could operate it.

Yay sales bot! Give that guy a raise!

Tech winThe first automation works in production

Ops momentA staging URL meets the whole company

Growing upThe first real security processes arrive

2019

Eventbrite / payments / WSS

The double life

Startup architect by day. Bachata champion by night.

Audience Republic expanded through ticketing integrations and a growing engineering team. Outside the office I won two WSS bachata categories and medalled in another — the other half of a life that work Slack mostly never saw.

A startup, a white panel van, and two championship wins in the same year.

Personal winTwo WSS bachata categories

Still on callEvent syncs do not observe championship week

Private archiveDance memories saved in notes to self

2020

pandemic / reliability / scale

Pandemic builder

The event industry stopped. We kept building.

COVID emptied calendars, but the platform and team continued to grow: hundreds of thousands of email sends, better monitoring and two people online testing deployments while Sydney stayed home.

It’s just you and me online — it’s a pretty quiet New Year’s.

The contextAn events company in a world without events

Scale704,000+ email sends monitored in real time

NYEA very small deployment party

2021

infrastructure / teaching chapter

Infrastructure + endings

Foundations, late nights and an end of an era

AWS networking and distributed systems got deeper — the kind of debugging that ends with tcpdump after midnight. My regular dance-teaching chapter also closed at the old Tropical Soul studio.

Such a fun last week of classes… end of an era.

Late nightsUDP multicast debugging after midnight

Classic JasonRunning on empty; need a coffee

PersonalThe regular teaching chapter closes

2022

scale / delegation / still debugging

Scaling the organisation

Learning what to hand over

Engineering needed structure that did not depend on me doing every kind of work. We hired dedicated management, deepened the technical partnerships and — inevitably — broke Emacs after a Homebrew update.

After running brew update I’ve now broken Emacs.

DelegationThe first dedicated engineering manager search

Holiday modeDocker and Postgres debugging on 30 December

UnresolvedStuck outside my apartment — long story

2023

integrations / incidents / culture

Code + culture builder

Culture became part of the architecture

More ticketing integrations, squad leaders and the occasional 2:30am incident. I was still in the code, but more of the job became teaching through failures, recognising people directly and making the team safer to operate in.

2:30am is not a fun time to be looking into this.

The long view25+ continuous years of professional coding

IncidentOnline when it matters

LeadershipChecking the people, not only the output

2024

performance / candour / principles

The honest year

A necessary, uncomfortable reset

Too much of the year disappeared into making legacy code perform. Over Christmas in New Zealand I wrote down the honest version. That reflection became clearer engineering principles and a reset for what came next.

I’ve spent a lot of 2024 just refactoring code — actually exhausting and depressing for me.

PerspectiveA strategic reset written from New Zealand

CodifiedEngineering design principles become explicit

Rare candourSaying out loud what the work was costing

2025

AI adoption / AM3 / codification

Culture codifier + early adopter

AI moved from private experiment to team practice

I tried Claude Code and multi-model review on my own work first, kept the examples that genuinely helped, then brought the useful parts to the team. Audience Manager 3 shipped after years of architectural decisions.

Try it privately. Keep what works. Make the useful part normal.

AI adoptionPersonal experiment before team standard

MilestoneAudience Manager 3 finally ships

OnboardingPrinciples from day one for every new hire

2026

architecture / sale / next things

Architect, delegator, still in the code

Still close enough to remove 397GB of dead indexes

The title shifted toward architecture, the tools became standard and Audience Republic was eventually sold. I still wanted the foundational work nobody notices — and a current folder full of Neuma, Lullafi and a gig agent for my sister.

397GB of dead indexes dropped. Invisible work; very visible relief.

Tech winFour tables lose 397GB of dead weight

Team standardAI-assisted engineering becomes normal

What nextNeuma, Lullafi and useful personal agents

the bit that survived every era

Stay close to the work.

Titles changed. Scale changed. The useful instinct did not: understand the real problem, make the complicated part feel simple, and keep enough of your hands on the machinery to know when the abstraction is lying.