Skip to content
Now building

Making AI work for the people doing the real work.

I build AI tools for professionals, small companies and NGOs, and write about using them well, from a part of the world the big platforms often leave out.

AI & Automation Engineer at Nile Automate 15 years across Juba, Nairobi and Kampala

15+

years in humanitarian communication across East Africa

7

products built, from job search to community video

31

testers in Kazi Board's first quiet week

100

verified job matches found for those testers

What I have built

Seven products, from job search to sportsbooks.

Notes from the build: what each product took, what broke, and what the people using it told me. Building with AI still means hours of instructing, testing and fixing, and I keep those hours visible.

Kazi Board home page
Job search In testing

Kazi Board

An AI job-search assistant for people who shouldn't have to check six job boards every night. It builds your CV from your real work history, matches you to verified openings and drafts applications for you to review. In its first quiet week, 31 testers created 12 CVs and got 100 verified matches.

Visit kaziboard.com

What I write about

Three subjects, one question about who gets in.

Every piece starts from the same place: someone trying to get real work done with tools that were built far away.

Local and open-source AI

Models that run on an ordinary laptop, offline and free, with nothing you type leaving your computer. Step-by-step guides for readers who cannot pay for a subscription or reach the big platforms.

AI for real work

What to automate first, what to leave alone, and why a person should check the work before anything goes out. One workflow run properly for two weeks teaches more than ten tried in an afternoon.

AI outside Silicon Valley

What these tools mean for people in Juba, Kampala and Nairobi, far from where they are made: payment walls, patchy connections, language, and who controls the tool.

From sitreps to systems

I build with AI and write down what I learn.

Crypto is where the building started. At Nile Capital and on Betzone I learned to scope a product, coordinate developers, work with government and launch something real in Juba. AI is where that building happens now.

The through-line has held across all three fields. Complex systems fail when information is badly communicated, half-understood, or late. AI makes that problem faster and larger. It also puts the fix within reach of a small team, or one person with a laptop.

The full story on the About page
  1. 2008 World Vision · Juba First field stories and situation reports, written for readers who had to act on them.
  2. 2017 REACH · Juba Assessments and briefing papers that fed humanitarian decisions.
  3. 2023 Nile Capital Crypto research, investor education, and building Betzone into a licensed business.
  4. 2026 Nile Automate AI & Automation Engineer. AI systems for small businesses and professionals, with a person checking the work.

Featured writing

A small list, chosen deliberately.

Three essays at a time, rotated every four to six weeks. All essays

Transitions

A toolbox, translated

The thesis piece. Why the skills developed across fifteen years of humanitarian work are almost always the exact skills the private sector is short on, and why the translation is a vocabulary problem rather than a capability problem.

Read
AI is a powerful tool, and it is still a tool.
From “My shift to building with AI”

Why this site exists

Slow writing for a fast field.

Every week brings a new model, a new benchmark and a new claim that everything has changed. What is often missing is the account of what it took: the hours, the failures, and the checks.

The habits I built in humanitarian work (checking sources, writing down what I do not yet understand, testing every claim against what people actually did) are the habits building with AI asks for, every hour. When the field moves fast, I answer by writing more carefully.

Start here

Pick the path that fits why you came.