The long way round
I did not take the straight line into AI, and I'm glad. Biology taught me precision, games taught me to build, and the rest taught me to trust the work. Here's the path.
- Earlier
Paying dues (yes, including KFC)
Before the lab and the models, there was customer-facing work, a stint at KFC included. It taught me speed, composure, and that the customer's problem is the only problem. Underrated training for shipping.
- 2020–2024
A scientist's training
B.Sc. in Molecular & Cellular Biology at Kenyatta University. Lab protocols taught me precision under pressure, a 98% sample-processing success rate, and the habit of trusting evidence over vibes.
- 2020–2022
Learning to build
Built Unity game prototypes in C# with AI behaviour scripting and Blender assets. This is where object-oriented thinking and systems design clicked.
- 2023–2024
Teaching the machines
Evaluated and trained LLMs at Outlier AI, grading RLHF outputs, annotating 200+ code snippets, ranking top 10% for QA accuracy.
- 2025–2026
Frontier work
LLM and video-annotation training for Turing and iMerit: structured frameworks that cut annotation ambiguity, and high-fidelity audits of generative media.
- Now
Building & giving back
Shipping agentic AI products, finishing a CS degree (University of the People), and helping people build safe superintelligence one piece at a time.
Like how that turned out? Let's add a chapter together.
I'm open for projects, training work, and workshops.