7 min read
What the Road Makes Possible
Why walk this road at all? Once the foundations are in place, one team can help a business know its customers, keep them, price fairly, understand big risks and work faster.
Writing
Notes from inside the build. What the engineering actually costs, where the received wisdom fails, and what I would do differently. Every example comes from a system that exists.
Stay in the loop
New posts on AI systems, engineering craft, and lessons from building in production. No spam. Unsubscribe anytime.
data science9leadership8ai agents7software engineering6ai2architecture2data engineering1mlops1open source1python1
Series
All series →Building a Data Team
8 of 8 published- 01How to Build a Data Science Function From Scratch: A Roadmap
- 02Who You Need on the Team
- 03Stage 1: Find Out What Is True
- 04Stage 2: Give Your Data One Safe Home
- 05Stage 3: Give the Work a Clear Path
- 06Stage 4: When People Push Back, Show Them
- 07Stage 5: Trust Your People, Then Let Go
- 08What the Road Makes Possible
Our Version
2 of 2 published- 01We Don't Have Rockets. We Have Towers. A Case for Building Africa's Data Centres
- 02Where the Towers Are. A Map of What Africa Already Has
2026
6 min read
Stage 5: Trust Your People, Then Let Go
The last step is the hardest: stop being the person everything depends on. How trust moved us from waiting for things to break to stopping problems early, with a 90 day plan.
6 min read
Stage 4: When People Push Back, Show Them
The technology was not the hardest part. People were. How we got through pushback and red tape, by showing value in action, and how a crisis became our best argument.
8 min read
Stage 3: Give the Work a Clear Path
It used to take about a year to get a model into use. Now it fits in a two week sprint. The change came from a shared workspace and one agreed path, not from cleverer people.
5 min read
Stage 2: Give Your Data One Safe Home
A model can only be as good as the data it learns from. How we brought scattered data into one checked and controlled home, and made security part of normal work.
5 min read
Stage 1: Find Out What Is True
The first stage of building a data function is not technology. Write down what you have and create a rhythm. When we did, we found models that were simply wrong.
5 min read
Who You Need on the Team
Who belongs on a data team, who to hire first, what to do when the budget runs out, why graduates matter, and where the team should report.
8 min read
Where the Towers Are. A Map of What Africa Already Has
Part one argued that Africa's cell towers could be a data centre spread across the land. This part asks the obvious next question: where are they? Using open, crowdsourced data on 565,000 mobile cells, it maps South Africa and the continent, finds that the towers follow the people and the roads, and shows why a map of people and a map of land tell two very different stories.
22 min read
We Don't Have Rockets. We Have Towers. A Case for Building Africa's Data Centres
Google, SpaceX and others are putting computers in orbit because power and cooling on Earth have run out of room. Africa has no space programme, but it has tens of thousands of cell towers that already have power, batteries, cooling, a fence and a link to the network. This post asks whether those towers can be a data centre spread across the land, and answers with the same graph and scoring method I used to send generators to towers during load shedding.
7 min read
How to Build a Data Science Function From Scratch: A Roadmap
A five stage roadmap for building a data science, data engineering or AI function from scratch: what you will find, what to do first, what pushback to expect, and how to know when to move on.
10 min read
The Practical Roadmap to Building With AI Agents, Part 6: From One Agent to an Agentic System
Loops, subagents, orchestration and routing, explained without the hype, plus the economics nobody mentions. Part six of six: what this site actually runs, what it deliberately does not, and why more agents is often the wrong answer.
11 min read
The Practical Roadmap to Building With AI Agents, Part 5: Making Agents Reliable Enough to Ship
Generation gives you code. Feedback loops turn it into engineering. Part five of six: the failures on this site where a green build was exactly the wrong goal, and how a bug becomes a permanent check nobody has to remember.
9 min read
The Practical Roadmap to Building With AI Agents, Part 4: Making Agents Safe Enough to Act
If your agent can use tools, what stops it doing something stupid? Part four of six: why prompt instructions are not a security control, the database rules that make bad outcomes impossible on this site, and the gaps I have not closed.
10 min read
The Practical Roadmap to Building With AI Agents, Part 3: Giving an Agent Knowledge
The assistant on this site answers from my actual writing. Part three of six: RAG explained without the jargon, the chunking and search settings I actually use, and three retrieval bugs where the model was fine and the information was not.
10 min read
The Practical Roadmap to Building With AI Agents, Part 2: Giving an Agent Hands
An agent that can only edit files makes you the courier between it and everything else. Tools and MCP fix that. Part two of six: what a tool actually is, what I connected, what it fixed in thirty seconds, and the permissions file I am not proud of.
11 min read
The Practical Roadmap to Building With AI Agents, Part 1: From Chat to Agent
Most people use AI like a vending machine. I used it like a senior engineer sitting next to me, and built this whole site that way without knowing Next.js. Part one of six: what actually separates a chat window from an agent, and how to start.
9 min read
The Modern Data Scientist: A Roadmap for the Age of AI Agents
The job hasn't disappeared. The mental model has changed. Here's what the roadmap looks like for data scientists and engineers entering (or navigating) the AI agent era, with a South African lens.