0° Equator

Nairobi · 1.29° S · 36.82° E

Field notes on intelligence that works in the physical world.

This is where I keep what I'm studying, testing and reaching for: robotics, physical AI, and the systems Kenya will need next. Written in public, as I learn.

SI = superintelligence. A direction, not a claim.

Threads I'm pulling

StudyingSince Oct 2026

Robotics and physical AI

I'm working through Modern Robotics by Kevin Lynch and Frank Park, from the foundations up: configuration space, degrees of freedom and rigid-body motion. The aim is to move from models on screens to machines that sense and act.

Personal R&D

Edge vision for drones (AODS)

A personal research project on onboard computer vision for anti-poaching, search and rescue and perimeter monitoring, where internet and GPS can't be relied on. Compressed detection models on small devices, sensor fallbacks, and a human in charge of anything consequential.

Prototypes

AI for health where specialists are scarce

Retinal screening with EyescanAI, demonstrated to senior ophthalmologists in Kenya, and clinical reasoning support for nurses (11th of 1,640 teams in the Kenya Clinical Reasoning Challenge).

2024

Seeing crises before they peak

Forecasting humanitarian crisis severity three months ahead with ACAPS and Omdena. I want to bring the same forecasting discipline to infrastructure and supply chains.

Exploring

The physical economy in Kenya

Kenya is known for services. I'm looking at what sits underneath: energy, infrastructure, manufacturing and logistics, and where AI and robotics fit in. Mostly conversations and learning for now.

Questions I'm sitting with

  1. What would a robot need to be genuinely useful on a Kenyan farm, road or factory floor, and what would it cost?

  2. Where does edge AI beat the cloud in places with weak connectivity?

  3. What does a large energy or infrastructure project need that nobody has built yet?

  4. How do you keep a person in charge of a fast system without slowing it down?

On my desk

  • Robotics

    Modern Robotics, Course 1: Foundations of Robot Motion (Lynch and Park, Northwestern)Since Oct 2026Studying

Notebook

  1. The study group has started

    A diverse community, across the world, one mission (working through Modern Robotics, Course 1 together). We check in daily on the Fellowship, the group's invite-only app, and meet every Wednesday, starting on 21 October.

  2. Starting robotics, and looking for study partners

    Why I'm adding robotics to machine learning, why I'm starting with Modern Robotics on 21 October, and how to join a study group of up to 7.

  3. Why kaime.si exists

    A public place to think about what comes next, including the parts that don't work out.

New entries land here first, then onLinkedIn

Where this is going

Long term, I want to do machine learning research that ends up working in the physical world, study further, and build ventures in Kenya's physical economy. This page will show the steps, including the ones that don't work.

Building something physical?

If you work on robotics, physical AI, energy or infrastructure in Africa, I'd like to hear what you're building.

Now · kaimenyi.me

That's where I'm heading. What I'm building today lives on kaimenyi.me.

Visit kaimenyi.me