Abdulhamid
Haidar

Open WhatsApp. Type a number. A math tutor texts back. That is the whole interface - and for more than 120,000 students across Jordan, India and Nigeria, it is school.

Abdulhamid Haidar runs Darsel, a nonprofit he founded and leads as CEO. Darsel is a math chatbot, but the word undersells the bet behind it. While most education technology assumes a student has an app, a charged tablet, and reliable home Wi-Fi, Darsel assumes almost none of that. It rides on the channels people in low-income communities already have and already trust: WhatsApp, Facebook Messenger, and ordinary SMS.

The numbers tell the story. In low-income countries, fewer than 20% of people have internet access - but more than 70% own a phone. Conventional edtech designs for the first number. Darsel designs for the second. Students practice curriculum-aligned math through plain messages, get hints and explanations, and work through a library of tens of thousands of questions in multiple languages. It is free for students and teachers, funded entirely by donors.

The experience is deliberately simple. Underneath, it is not. The chatbot runs AI-powered adaptive-learning algorithms that map each student's individual skills and gaps, then serve the right question at the right moment to push them toward proficiency. Simple on the surface, busy underneath - a calm front desk hiding a busy back room.

Spring 2020. The pandemic shut schools worldwide, and roughly one billion students lost access to education overnight. Haidar was midway through his MBA at Stanford, on a joint program with Harvard Kennedy School. His spring break plans evaporated. So he started building.

The seed had been planted on an earlier trip to Syria, where he watched how hard it was just to get online. When the world rushed to Zoom classrooms, he kept thinking about everyone Zoom would leave behind. "The digital divide was at the top of my mind," he has said, "as I had recently returned from a trip to Syria where internet access was a challenge."

He spent weeks during lockdown experimenting with a rough chatbot prototype. The instinct that texting could carry real learning hardened into conviction. "I felt I had landed on a solution that could actually make a massive dent in global education." The early pilot in Jordan was modest and convincing at once: about 100 students across two public schools answered more than 10,000 math questions.

He had planned to return to McKinsey after graduating. He went all in on Darsel instead.

To understand the choice, start with the dinner table. Haidar is Syrian-American, born into a family of educators - four of his grandparents were professors or teachers. His father taught him advanced math young, sometimes during walks to school, the way his own grandfather had once tutored him on the same kind of walks. In his family, education was not a ladder to climb. It was who they were, a piece of homeland carried wherever they went.

He grew up split across two places - summers in Syria, winters in the United Arab Emirates. Then came MIT, where he studied mathematics, computer science and economics. He started out aimed at software engineering, then pulled toward policy. After graduating he spent three years at McKinsey working on economic development and public policy, much of it with Middle East governments on issues including education. On the side, he volunteered on refugee-education projects in Jordan and Lebanon.

Syria's 2011 civil war reshaped how he thought about impact - away from individual fixes and toward systems and institutions. The throughline from a grandfather's walk-to-school tutoring to a chatbot tutoring 120,000 kids is not subtle. He simply found a way to do at scale what his family had always done at the kitchen table.

Darsel runs in Jordan, India and Nigeria, and Haidar is blunt that the same product cannot simply be copy-pasted across them. Each country means a different curriculum, different languages, different partners, and a different classroom reality on the ground. He describes Darsel as "a solution that is global in some ways and local in others" - the engine is shared, the content and the fit are not.

That balance is the hard part of the work, and the reason it scales without flattening. The chatbot partners with public school systems and teachers rather than going around them. Teachers get data insights into where their students are stuck, which turns a texting tool into something a classroom can actually steer.

The recognition followed the results. Darsel went through Y Combinator's Winter 2022 batch, won a Tools Competition Growth Prize in 2023, and became a finalist for the WISE Prize for Education. It is backed by funders including the Mulago Foundation and Fast Forward. In 2025, Haidar was ranked eighth in a Gen Matrix ranking of Gen AI leaders.

The genius of Darsel is what it refuses to do. It does not stream video. It does not require a login dance, a download, or a fast connection. A student opens a messaging app they already use to talk to family and starts answering math questions. The cost to participate is close to zero, which is exactly the point in homes where a megabyte is a budgeting decision.

Behind that plain text exchange, the adaptive engine is doing the heavy lifting. It assesses what a student knows, finds the specific skill they are missing, and feeds the next question accordingly. Get something wrong and the bot offers hints and explanations rather than a red X. The library spans tens of thousands of curriculum-aligned questions, localized into the languages and standards of each country it serves. The aim is not engagement for its own sake - it is proficiency, measured one answer at a time.

Teachers are not bypassed; they are equipped. Darsel hands them data on where their students are getting stuck, turning a personal texting habit into a classroom signal. A teacher can see the gap before it becomes a failing grade. That feedback loop is part of why partners keep renewing - the tool makes the human in the room more effective rather than trying to replace them.

It is a quietly radical inversion of how most software gets built. The usual instinct is to ship the richest experience the best hardware can handle, then trim downward for everyone else. Darsel started at the bottom of the bandwidth ladder and built up. The constraint was the feature.

The stated goal is to positively and meaningfully impact 100 million students by 2030, directly or indirectly. It is the kind of number that sounds like a pitch deck until you remember the starting point was 100 kids in two Jordanian schools, and the line since then has only pointed up and to the right.

Increasingly his attention is moving earlier in the school journey. "Learning gaps become significant in primary school," he notes - which is the argument for reaching younger students before small gaps compound into permanent ones. The instinct is consistent with everything else about Darsel: get there early, get there cheap, get there over the channel the student already has in their pocket.

The framing he likes best is the simplest. The idea, he says, "came from the global wave of school closures that occurred at the very beginning of the Covid-19 pandemic, when 1 billion students lost access to education." Darsel is his answer to that sentence - one math question at a time, in a chat window, for a student a continent away.