Ventures
Six companies built and running, and one exit. Every one of them started from a problem I could name precisely. They are also where I find out what works before I write about it.
Disruptive commerce platform — 0% cost and 0% commission for local shop owners.
A disruptive commerce platform for local shops: storefronts, ordering, payments, and delivery orchestration — free for shop owners, monetized through disruptive omnichannel revenue streams. Built AI-first: shop onboarding and customer ordering run on own trained and developed AI models, and the platform itself was built and scaled with AI tooling.
The operator stack for Introducing Brokers — grow the book, then run it.
Most IB software only handles administration. IBHQ starts a step earlier, with AI-powered marketing and outreach that changes how an introducing broker actually builds a business: campaigns and content generated and published for them, automated multilingual outreach and onboarding bots that qualify and convert leads without a marketing team. Underneath sits the operating layer — client onboarding, rebate and sub-IB settlement, signal forwarding and copy-trading in one control plane, Telegram automation, and bilingual (EN/AR) brand sites. The result is a one-person IB able to run like a firm.
Booking software for salons, barbers and clinics.
Online booking built for shops that run on appointments: SMS-code confirmation, a calendar with a column per staff member, client records, automated reminders, and a waitlist that fills cancellations — presented in each client's own language. Migration from incumbent systems is done for the owner, free, during the pilot, because switching cost is the real reason small businesses stay on software they dislike.
The platform that treats language teachers as business owners, not gig workers.
Gamified apps optimise for streaks; fluency comes from a person who corrects you. KoloLingo gives that person a business. A teacher gets their own academy, not a profile on someone else's marketplace: their prices, their curriculum, group classrooms, recorded courses, downloadable materials and lesson recordings they can sell, and a growth path as their reputation compounds. Commission is 15% flat, calculated after payment processing and never rising with tier, and only 5% on students the teacher brings themselves, against the 18-33% sliding scales common elsewhere. Language schools run their whole operation on it: invite link, their own teachers and students, the revenue share they agreed collected and paid out automatically, an owner dashboard and a shared material library. AI does the unglamorous half of teaching, cutting preparation and admin so the hour with the student is the work. For learners there are level tests, structured courses across more than thirty languages, and instant lessons: no booking, no calendar, see who is free and be talking in under sixty seconds.
Trade what you see, not what you think — a method, an algo engine, and the tools to run both.
A bilingual (EN/AR) trading platform built on one idea: no indicators between you and the candle. It carries MT ALGO, an automated engine run with a single parameter set across unrelated instruments and timeframes with the entire track record published rather than a curated month; a live Markets Desk terminal per instrument with levels, alerts, replay and the release calendar; a chart explainer that reads a chart back to you; plus the academy, live sessions, channels and leaderboard around them. Payments run on cards and on USDT/USDC rails, because conventional rails are often not an option for the audience.
A clear method, real tools, and people who have walked the road.
The sister platform to MT Trading, also bilingual (EN/AR), aimed at traders who want the tooling as much as the teaching: DM Algo Ultra, an automated strategy across gold, world indices, oil, crypto and currencies with every trade published month by month; the Markets Desk terminal; a chart explainer and chart clinic; signal channels, academy and community. Funded through a partner-broker model so the method and the tools stay free to the trader, which also makes the economics honest: if the tools do not get used, nobody gets paid.
Ongoing · Builder · Quant
An adaptive quant system, plus the analytics stack around it.
A machine-learning quant system that keeps re-fitting itself to changing market regimes instead of assuming the conditions it was trained on still hold — the hard part is never the signal, it is knowing when the signal has stopped working. Around it sits a full analytics stack: custom indicators, automated execution, backtesting with honest cost and slippage assumptions, and risk tooling. It is also where I test, with my own money at stake, whether the AI claims I write about survive contact with a market that does not care about the demo.
SaaS ordering & loyalty platform for hospitality.
B2B2C ordering and loyalty SaaS for restaurants: product vision, pricing, and go-to-market — built with a near-shore engineering organization in India on AWS and DigitalOcean.