Mohammad Abu Sheikh, founder and CEO of Cntxt AI: Each week, My Morning Routine looks at how a successful member of the community starts their day — and then throws in a couple of random business questions just for fun. Speaking to us this week is Mohammad Abu Sheikh (LinkedIn), founder and CEO of Cntxt AI, as well as several other ventures he has since exited, and managing partner of Smpl AI. Edited excerpts from our conversation:

My name is Mohammad Abu Sheikh; I’m the CEO and founder of Cntxt, and I just like to build things. I also have a fund, Smpl, that invests in seed-stage AI startups globally, which is self-funded. I started working when I was 12, and I started my first business when I was 16. It was an intellectual hub back in Jordan — I wanted to build a place for youth to sit together and share thoughts and ideas, and I added coffee to the mix to make it more fun. It grew to five or six branches across Jordan, and it’s opening in other countries soon, though I’m not managing that anymore.

We had a very big kitchen in Kitab and we didn’t use it at all, so I decided to start a cloud kitchen business long before they became a hype. That was my second business. I ended up opening another cloud kitchen in Dubai, and then sold the business later to a family office.

When I was supposed to start university, there was a big question mark for me, because I knew that a lot of people — most of my employees at Kitab — do not work in the field that they spend years studying. My family wanted me to go, so I went, and I was building my cloud kitchen business at the same time, which had a user interface and ordering application, as well as several other backend applications. One day, I showed my professor my app, and he said it’s great. I later showed him the code for my app and asked for his opinion on it, and he said the code is sh*t. I had to ask him if he ever built an app or a software of any kind, and he said no. That’s who was teaching us to code.

That day was the last day I attended university — I withdrew my file from administration, and I just continued to work on my app. I began to think there’s a fundamental problem with education, and I wanted to change the concept in people’s minds that education takes place in schools — I believe schools can school you, but they don’t necessarily educate you. I decided to build something that was interactive, and we started by creating an LMS / CMS system that allowed students to go in and take a placement test, and based on the results of the test, they’re taken through a learning pathway that is personalized to their preferred method of learning, whether that’s audio-based, visual, or through gamified experiences. We gained a lot of traction, and we started working with around 80 schools, and with the government here through the Mohammed bin Rashid Foundation, and in Saudi Arabia and Qatar as well. We sold the business after 18 months.

I wanted to gain more [market share] and help more people learn faster, which inspired me to create an audiobook summary business — Lamha — and we sold that as well after reaching 150k active and paying users.

That brought me to my journey with AI. A friend of mine who leads a semi-government entity in Abu Dhabi wanted to build the first Arabic large language model (LLM) and they were failing miserably, so they came to me to help out since we’ve had experience building a lot of things in the tech space, and that’s how Cntxt started. We started working with G42 and the Technology Innovation Institute (TII), and directly with the government. We also started another business, LocAI, to build applications on top of the large language models. LocAI began to gain a lot of traction among public entities like the department of municipalities and transport, ADGM, and others.

We had no clue how to build a large language model, but we knew how to build things very well. We approached it like a problem statement, and dissected it into three different parts: data, compute, and science. The algorithms were working pretty well because they were world-class, the compute was Nvidia — also world-class — so the only problem was the data problem. That’s what required us to dig deep.

I always prioritize from the day before. I make sure I know what I’m doing the next day, and I like to time block, so for instance, Mondays are for product, Tuesdays are for tech, Thursdays are for marketing, Wednesdays are for funding, and so on. But 95% of the time, this doesn’t work; the chief thing I do in my day is problem solving, and then the other portion of my job is building out towards the vision and staying in sync with the market to make sure that we’re building things that are still relevant.

My morning routine is nothing special — I start my day by showering and praying, and I get to work as early as possible to gain as much lead time as I can.

In terms of personal hobbies, I enjoy padel and reading and listening to audiobooks, because I commute a lot between Dubai and Abu Dhabi. I’m reading a book called Everything is Predictable, by Tom Chivers. It basically says you can — scientifically — predict anything just by observing and understanding and having the data points, but it also ends with the thought that it’s still not possible to predict everything — because sh*t happens. I also really like The Hard Thing About Hard Things by Ben Horowitz, and would also recommend, if you’re building anything in consumer tech, The Cold Start Problem by Andrew Chen, and Contagious if you want to read about marketing and business growth.

A very honest piece of advice that resonated with me, which I got from one of my mentors is: if you’re gonna eat sh*t, don’t nibble. It just means if you’re going to do something hard, don’t dread it or think about it too much, just do it.

Tags: