About Wazomind

Wazomind is the meeting co-pilot that helps while the call is live — out in the open. It was built end to end by one person who spent seven years in the meetings that inspired it.

George Muchiri Wanjii

Founder, Wazomind · Nairobi, Kenya · works in English & Swahili

wanjiippc.co.keGitHub

George Muchiri Wanjii is a developer and digital strategist based in Nairobi, Kenya, with seven years across web development, data analysis, and growth.

He has led digital and marketing functions at AfricaSokoni, where systematic keyword and on-page work lifted organic traffic 35%, and at Quatrix Global, where auditing campaign logic for flawed assumptions cut cost per lead by 22%. At The V Bar he led acquisition for Nairobi's first app-based delivery platform, scaling new users past 40% and lifting conversion 30% through iterative, data-validated experiments. Alongside those roles he has built and optimised web applications for clients across Africa, Europe, and the US — React front ends and API integrations, custom PHP functionality, B2B commerce platforms, and analytics instrumentation in GA4 and Looker Studio.

Two threads run through all of it: a habit of decomposing problems into step-by-step logic, and a discipline of auditing work for the gap nobody caught yet.

I spent seven years in other people's meetings — client calls, leadership reviews, the moment a director asks for a number and waits. I kept reaching for a tool that would help me then, not in a summary email afterwards. The ones that existed either arrived too late or worked by hiding from everyone else in the room. So I built the one I wanted: it helps while the question is still in the air, and it does it in the open. — George Muchiri Wanjii, founder

Wazomind came out of the meetings. Years of client calls and leadership reviews — distilling performance data into reasoning he had to defend on the spot — made the gap obvious. Transcription tools had become excellent at telling you what happened after a call ended. Nothing helped while the conversation was still live, when help would actually change the outcome. The one category attempting real-time assistance had chosen to do it covertly, hiding from other participants.

He took the harder path on both counts. Wazomind assists during the call and does it transparently: a recording indicator that can't be disabled, no undetectable mode, and no bot that joins on your behalf. It runs local-first — transcription and coaching can operate entirely on your own machine, with no audio leaving the device. Both choices cost more engineering effort than the alternatives. Both were deliberate, on the reasoning that a tool a company can actually approve is worth more than one that goes viral and gets banned.

He built the product end to end: a Rust desktop shell with dual-channel audio capture, real-time transcription with speaker separation, a retrieval layer grounding every suggestion in the user's own documents, and a metered billing backend — shipped to production on infrastructure he runs himself.

George holds a Bachelor of Business Management from Inoorero University. He works in English and Swahili.