The person behind the products
How I became
this kind of engineer.
A Computer Science foundation at Meru University, practical enterprise work, international teams and the responsibility of building products myself. The progression matters more than a list of titles.
02 / How I got here
Increasingly difficult problems.
The same curiosity.
From enterprise workflows to scientific software to products of my own. Each chapter changed how I think about engineering.
201901
A foundation in real systems.
Enterprise recruitment software and remote client assignments brought frontend, backend and data modelling together. At EnigmaScore and on SpotADev work, I learned to think through the whole workflow—with attention to the whole system.
BSc Computer Science · Meru University of Science and Technology. Exact study/employment overlap is not documented.
2019–2102
AI, before the current wave.
My early machine-learning work included Android OCR and training Keras models on EMNIST in Google Colab. By 2021, my CV documented GPT-3 beta experience; later client work for Prospect 33 applied language models to interview and question scoring.
TensorFlow · Keras · Python · GPT-3 beta. Early use is documented, rather than claimed from launch day.
202103
Taking ownership of the product.
I started SelfAwareTech as a concurrent independent product studio. Building Stok meant owning the unglamorous parts too: inventory rules, persistence, mobile releases and what happens when the network disappears.
Founder & independent engineer · February 2021 onward.
2023–2604
Engineering for scientific complexity.
With the distributed Cell Collective team, I worked on scientific modelling software supporting thousands of users. Shared tooling, simulation workflows and release quality became as important as the individual feature.
20+ production features · 10,000+ platform users · approximately 40% faster builds. CV-reported outcomes.
2025–2605
Getting closer to the user.
Nyumba brought service providers and customers into the same product. Its Android and web work made discovery, requests, trust and onboarding concrete engineering problems. A request can now survive the transition through sign-in.
Reported milestone: 1,000+ Play downloads and 120+ providers · three-city pilot.
Now06
Useful intelligence, explicit boundaries.
Tasko brings company knowledge, retrieval and department workflows together with permissions and reviewed actions. Haulio turns freight demand and available truck capacity into an explicit domain model. Both deepen the same practice: understand the problem, then build the system around it.
Active applied AI and logistics work · public beta and ongoing production iteration.