Everything I've built & worked on

projects

• LORE (2026–present): an organisational knowledge-transfer platform, built to capture what an experienced employee actually knows before they leave — from what they do rather than how they'd explain themselves if asked. Built and tested on Firebase and vanilla JavaScript. An early version paused after user research showed the HR/L&D workflow I'd assumed didn't match how these teams actually work. Back in active development since, after a proper diagnosis of what the knowledge-extraction pipeline was actually missing — though I'm no longer treating it as a startup to raise money for. What LORE and the wider Learning Oven work taught me is that most people and companies don't yet see this as something worth commercial investment. Right now it's closer to research: working out, properly, whether the underlying idea actually holds up. Revisiting it this closely is also what sent me back to write out the fuller, nine-year history behind the underlying idea — I'd lost track of my own throughline and needed to see it laid out plainly again.
• SYD — Synchronized Yield Directive (2026): a self-improvement RPG that runs entirely in the browser — vanilla JavaScript, no framework, no account — so it works on any device, including low-end ones. Effort across five attributes (Strength, Intelligence, Agility, Endurance, Charisma) compounds into a derived Luck stat, and a rolling 30-day "Behavioral Trace" means the system responds to what you've actually been doing, not what you say you'll do. Live and open source (AGPL-3.0) at syd-protocol.github.io/terminal. This is where the idea behind LORE actually started — the rest of this list traces how it got there.
• /signal (2026): a manhwa-style serialised story built to popularise the system behind SYD, so other people could see and use the same idea. I stopped once I realised I was building it more for myself than for an audience — in hindsight, that instinct was really me trying to see a concept more clearly, one that later became LORE. Live at syd-protocol.github.io/signal.
• Slop Runner (2026): a web game for training professional pattern recognition, built right after SYD to test another angle of the same underlying idea — partly to see if the concept could even work as an app at all. Building it is what made the full shape of the idea click: I merged it with SYD, picked the customer segment I thought would pay best (B2B), and rebuilt the concept to fit that segment. I now think that's exactly why LORE first stumbled — I'd built an idea to fit a chosen segment, instead of starting from a real problem that segment actually had. Live at slop-runner.github.io/game.
• /loop (2026): the fuller evolution of SYD into a top-down pixel-art life-and-career RPG — a separate, ongoing thread rather than part of LORE's line directly. Live at syd-protocol.github.io/loop.
• MangaTracker (2026): a browser extension and companion web app for tracking manhwa, manga, manhua, and anime across any reading site, with Firebase-backed sync so a library stays consistent across devices. I built it to track my own manga list, and still use it that way. Live at https://osioke.github.io/mangatracker-web/.
• Growth Clinic & Community Staples (2019–2025): an open, distributed accelerator and incubator for communities, and its sister publication on the science of community management. Growth Clinic secured $10,000 in non-equity funding; Community Staples guided roughly 30 people through a custom training programme. Wound down at the end of 2025 — the model had drifted from what actually worked at Devcenter (placing people in real paid work, taking a share of the value) toward selling training directly to individuals.
• Job Hunter Assistant, formerly JobRelay (Apr–Dec 2025): a browser extension — plus a companion web app at jobhunterui.github.io — to make job hunting less repetitive: saving listings from any job board, generating AI-assisted, role-tailored CVs and cover letters, and tracking applications. Rebuilt from an earlier, more ambitious version (JobRelay, started April 2025) into a leaner one focused on what people actually used. Sunsetted: most of what it did well turned out to be nice-to-haves rather than must-haves, so growing it would have taken a lot of marketing and growth work on top of real product problems — payment integration issues, and a roadmap that kept shifting because building fast with AI made everything feel achievable. When the people who tried it weren't sticking around, it was clear that scaling it meant working a lot harder for not very many users — not the trade-off I want to make.
• What Is It Gameshow (Jul–Nov 2025, contract): an interactive gameshow application for a live streamer, built with Electron. Scoped, built, and delivered independently to brief as a paid engagement.
• Shop (2025): a full-stack inventory and sales-management system I built for my mum's convenience store — role-based access, real-time inventory tracking, credit management, and sales reporting. Built solo with Firebase. She hasn't ended up using it day to day, but it's a solid, complete proof of taking AI-assisted building from idea to something a real business could run on. Live at https://github.com/osioke/shop.
• Gani Nigeria (2016–2025): a digital museum documenting Nigeria's culture and beauty. Started in 2016 while I was working at eHealth Systems Africa in Sokoto, as a way to share the North of Nigeria — the place I was actually working in — the way pictures and stories about it usually didn't. Continued through 2025. I stepped back after realising the fuller story I wanted to tell was more layered — culturally and politically — than I could do justice to without dedicated funding and time, and after my own priorities shifted when things got financially tight in 2024–2025, as Growth Clinic and Community Staples were winding down.
• Frintern (2018): a platform matching people to tech skill paths suited to their natural affinities, with a personalisation algorithm that hit over 90% accuracy across 1,000+ career matches, and a microlearning system delivered through WhatsApp and Slack. In a March–May 2018 pilot, participants learned a skill and applied it in 2–3 hours, down from a 20-hour baseline. Used to deliver paid training to African Fintech Foundry's Accelerator cohort and to Maliyo Games ahead of a product launch. A cohort from this era also built Owmi, a chatbot-and-app platform connecting Lagos residents to trusted, treated water sources, tested directly with real residents — not continued, since it needed ongoing in-person operational work at a point when the direction had shifted toward digital products. Frintern (later continued as techstarta.com) is now sunsetted — I left it in my co-founder Jide Owosakin's hands to go rework the same underlying idea into what became Growth Clinic. I didn't have the technical ability I have now to keep building it myself, back before AI-assisted coding existed, and we'd never found the right technical co-founder fit either.
• Learn by Doing (2016–2017): the first deliberate attempt to build a system around that same observation — combining real project work, tasks broken into manageable pieces, and one-on-one apprenticeship mentoring. Built at Devcenter, as a project for Devcenter, before I left to start Frintern. Failed honestly and quickly: real enthusiasm at the start, a consistent drop-off after roughly two weeks, and incentives that didn't fix it.
• Disease Info (2016–2017): a health-information progressive web app with an open API, built with a distributed team at Devcenter to make verified information on diseases in Nigeria accessible across media channels, as covered by TechCabal at the time. It's a Devcenter product now — I left it behind when I left Devcenter, so it's no longer mine to maintain.

experience

A lot of these roles are short because of fit, not commitment. How I think and work doesn't sit neatly inside how most companies here operate, and it took a while to find somewhere that did. Discourse (CDCK Inc. — the company behind the open-source Discourse forum software used by thousands of organisations worldwide) was that place; I stayed three years and did some of my best work there.

I left in 2022 to commit fully to Growth Clinic, which I'd already been running part-time alongside the Discourse job. The timing felt right — community-based organisations were picking up real momentum then, riding the same wave as the crypto/Web3 world, and I thought I'd finally find people who understood what I was building. I didn't. The funding I'd raised and my own savings ran out before I found the right customers, and I eventually saw the actual problem: a mismatch between who I was selling to and what I'd built, not a market that didn't exist. Closing Growth Clinic sent me back into a full-time job search, which is what built Job Hunter Assistant, which in turn led to LORE — which I'm treating less as a product to commercialise and more as an open research question worth working out properly, on its own terms.

By the numbers, across these roles: 50,000+ community members served at Discourse, 2,500+ grown from scratch at Devcenter, 15% month-on-month sales growth at RelianceHMO, and a matching algorithm at over 90% accuracy at Frintern.

A bit more on the ones worth a proper mention, then the full list for reference.

• Discourse (Nov 2019–Apr 2022) — Community Advocate & Product R&D Lead for a platform serving over 50,000 members. Shipped a Salesforce integration plugin and a points-based gamification system, ran the Support Experts Programme turning power users into advocates, and built the enterprise onboarding audit framework used with clients.
• RelianceHMO (May–Sep 2019) — Team Lead, Products & Online Retail. Delivered 15% month-on-month growth in retail insurance sales, built a feature prioritisation framework adopted across departments, and led a full site redesign off the back of a customer journey audit.
• Devcenter (Nov 2016–Dec 2017) — Community & Product Manager. Grew the developer community from 100 to over 2,500 members in a year, generating over $6,000 in revenue through partnerships and sponsorships (GitHub, Flutterwave, Stickermule, Ingressive, Leadspace). Conceived and led development of Gigson.co, a freelance marketplace that became the company's main revenue product — still live today, though it's long since evolved past what I originally built and isn't something I manage any more.

publications

• The Learning Oven — the fuller write-up of the nine-year history behind LORE and everything before it: Devcenter, Frintern, Growth Clinic, and what actually tied them together.
• From Knowledge to Judgement (2025) — the whitepaper underpinning LORE's approach to extracting tacit, hard-to-document expertise.
• Building Communities into Economies (2019) — open-access book distilling what watching real Lagos communities (a farmers' market, a taco restaurant, a dance studio) turn shared interest into shared economic value taught me about designing systems around that pattern.
• In Lagos, Nigeria, A Farmers' Market That Sells All Week (2025) — published by the Platform Cooperativism Consortium, a global hub for research and advocacy around cooperative, community-owned digital economies, affiliated with MIT, Harvard's Berkman Klein Center, and the Oxford Internet Institute. A case study on TKD Farms Market: how a rotation system creates equitable access for 185 vendors, a shared customer database extends sales beyond market days, and democratic vendor meetings solve collective problems. Getting published by a platform with that kind of reach and credibility, on work I taught myself rather than studied formally, is one I'm genuinely proud of.

talks & essays

• 4 Steps to Community Growth (essay & framework), 2022
• Improving Your Community's Experience with an Experience Map (essay & framework), 2022
• Identity and Digital Spaces: A Cooperative Experience — RSD11 Symposium, 2022
• A Growth and Innovation System (framework & essay), 2019
• 3 Steps to Innovate (framework), 2019
• Enabling the Nigerian Informal Sector with Financial Services (essay), 2018

education