Chapter 05
The Convergence
2025–Present
The developer and the PM finally merged — managing pharma capital projects while building an AI agent ecosystem from the same instinct that started everything.
Auckland
The first thing I noticed was the pace. In Singapore, everything moves fast — work fast, speak fast, eat fast, walk fast. The city runs around the clock and you match its rhythm or fall behind. Auckland settles down at five.
Shops close. Streets empty. People go home. Coming from thirteen years of that constant motion, the quiet felt almost disorienting at first — like something was wrong.
Nothing was wrong. New Zealand just values different things. Work-life balance is not a buzzword here — it is the actual expectation. People finish work and go to the beach, which is twenty minutes from the office, not two hours.
Beaches, tramping tracks, and camping grounds sit within arm's reach of the city. People blend into nature regardless of age — families hiking the same trails as retirees, teenagers camping in the same parks as weekend tramping clubs. I moved to New Zealand before I had a job here. It was the dream destination, and I wanted to see if the reality matched.
It did.
Douglas Pharmaceuticals
Douglas was looking for a Project Manager in Digital Transformation. The job description read like my career in reverse — digital systems, cross-functional coordination, stakeholder management, capital project delivery. It fit like a jigsaw piece, even though I had zero pharmaceutical experience.
The first few weeks were about learning the Douglas way. Every pharmaceutical manufacturer operates under GxP regulation, and Douglas takes that seriously. There is literally an SOP for anything you can think of — how to manage documents, how to assess risk, how to validate systems, how to handle change. Getting to know Operational Technology was its own education.
My first project was the migration of an existing legacy Schneider BMS system to a modern SCADA platform. Building management systems control everything in a pharma facility — temperature, humidity, pressure differentials, clean room environments. The existing system worked, but it was ageing and isolated. A modern SCADA platform would centralise monitoring, improve alarm management, enable real-time data trending, and unlock integrations with other facility systems that the old architecture could never support.
This is the kind of digital transformation that does not make headlines but shapes an organisation for the next ten to twenty years. Replace a decades-old system with one that talks to everything else, and every project after it becomes easier.
The pharma environment uses everything I have learned. The developer in me reads system architecture diagrams and spots where the risks hide. The PM in me coordinates across QA, IT, manufacturing, and operations — translating between compliance concerns and implementation constraints. The team-builder in me knows how to get cross-functional groups pointing in the same direction.
Every chapter contributed something. This role uses all of it.
The second brain
A project manager's day is meetings, decisions, documents, communications, and context-switching. Most of that work involves gathering information, synthesising it, and acting on it. The tools most companies use handle one or two of those tasks well — a document manager here, a task tracker there — but they are not interconnected. They do not solve the problem collectively.
I started small. Meeting minutes are a repeated workflow — the same structure, the same sections, the same formatting, every week across multiple projects. Agentic AI had reached a stage where it could reliably handle that kind of structured task. So I tried it.
Then I tried something harder: project activity tracking. Then process lookups. Each new task worked well enough to try the next one. The developer in me woke up.
What started as a time-saving experiment became a full system. A second brain — a knowledge engine that remembers every decision, every meeting, every correction across every project and makes it all searchable and actionable.
Agents that prepare meeting briefs by pulling relevant history, open risks, and pending decisions. A communication drafter that adapts tone based on the recipient. A process assistant grounded in the actual organisational SOPs, not generic PM advice.
The interesting part is not the AI. It is that every design decision draws directly from PM experience. The agents are narrow because general-purpose tools fail in practice. They show their reasoning because trust requires transparency.
Sixteen components now — three orchestrators for the complex multi-step workflows, thirteen specialised skills for everything else. A local-first RAG engine with a four-tier memory system. Bidirectional sync with the project databases. Shared protocols for quality control, session management, and safety guards that prevent the system from writing anything without review.
It is a personal project, built outside of work hours. But it exists because the developer who started this journey in a tiny Singapore studio thirteen years ago is still the same person managing capital projects in a pharmaceutical factory. The tools changed. The instinct did not — see a problem, build something that helps.
What I carried forward
I did not plan this arc. Moving from code to management felt like leaving something behind. Moving countries felt like starting over. Each chapter seemed like a departure while I was living it.
Looking back, nothing was left behind. The code gave me the instinct to build. The bridge gave me the habit of translating between groups who speak different languages. The scale taught me that growing a team matters more than growing a product. The systems showed me that complexity is manageable if you hold the thread.
Now all of those threads run at once — regulated projects during the day, agent systems at night, and the slow Pacific light of Auckland in between. The developer and the PM finally work on the same problem. That is the convergence.