Building a 19-year AI career on hunger, not hype

From Sakhr in the 90s to Director of Applied AI at Microsoft, Eslam Kamal’s career runs against almost every cliché about Big Tech hiring. He didn’t go to an international school. He didn’t feel ready when he applied. He got rejected the first time. He’s convinced impostor syndrome never fully leaves — and he thinks that’s fine.

If you asked Eslam Kamal at 22 whether he’d eventually run an applied AI org at Microsoft, he would have said no. Not out of false modesty — he genuinely didn’t feel ready. Nineteen years later, that feeling hasn’t fully disappeared. What changed isn’t the absence of doubt; it’s the discovery that doubt and hunger can travel together.

His path was built on small bets that compounded: an early childhood obsession with computers in Saudi Arabia, a Computer Science degree at Ain Shams in Cairo, a first job at Sakhr — one of the early Arab software pioneers that bet on Arabic NLP decades before LLMs — and a 5.5-year Master’s done in evenings while working full time.

A choice that decided everything: love over trend

When Eslam picked Machine Learning around 2007, it wasn’t a hot field. Most of the people working in it were academics. Junior engineers didn’t get hired into NLP teams the way they do today. What pulled him in wasn’t the salary or the trend — it was the Arabic language. He found machine translation and morphological parsing of Arabic genuinely beautiful, and that pull was enough.

That decision turned out to be the bedrock of everything that followed. Not because ML eventually became trendy, but because loving the topic was the only thing that made the years of late-night reading sustainable. He says it bluntly: if you don’t love what you’re studying, you can’t put in the hours the field demands — and the field demands a lot.

The first Meta rejection (and what it taught him)

Eslam didn’t apply to Meta the first time because he felt ready. He applied because his wife pushed him. He got rejected. He applied again later. He got in. He says the only difference between most candidates who get hired at Big Tech and those who don’t isn’t talent — it’s the willingness to apply again after the first “no.”

Impostor syndrome, he insists, never goes away. What you build instead is a quiet confidence in one thing: not that you know everything, but that you can learn anything. That’s the only durable skill in tech — and it’s a skill, not a personality trait.

The Arab education advantage no one talks about

Eslam now raises his kids in Seattle. He sees their school life is easier in many ways — but he also sees what they’re not getting: the friction that builds self-reliance. When information is handed to you, you don’t build the muscle of going to find it. Arab graduates, he argues, often arrive with a quieter superpower: they’re used to figuring things out themselves.

In a field that reinvents itself every 18 months, that habit is worth more than any specific framework. Resourcefulness compounds.

What he actually looks for when hiring

As a hiring manager in applied AI, Eslam reviews CVs and runs final interviews. He’s not impressed by years — he’s suspicious of them. A new graduate with two clean side projects, real statistics fundamentals, and the ability to explain their decisions often beats a 10-year veteran who’s done the same thing for a decade. The market is supply and demand, and Big Tech’s asking price is “prove you can ship something real, then tell me how you thought about it.”

Two practical signals he weighs heavily:

1. Applied projects, not just coursework. If your resume claims you know ML, he wants to see a small project that proves the theory was translated into code. Personal, freelance, hackathon — doesn’t matter, as long as it exists.

2. The ability to tell the story. In interviews, you’ll get behavioral questions: “tell me about a time when…” If you don’t have history, you can’t answer. Even one or two years of real experience — anywhere — gives you something to talk about. That’s the bigger reason Big Tech sometimes prefers an engineer with two years of solid local experience over one with ten years of stagnation.

Vibe coding, AI, and the carpenter analogy

On AI coding tools, Eslam offers an analogy. A carpenter’s job was never to swing a hammer. The hammer got replaced by a nail gun. The carpenter’s job is to build beautiful, useful furniture. Same with engineers: your job was never to type code. It was to design products, understand customers, and make good decisions.

If you treat AI as a junior developer you’re code-reviewing, it accelerates you. If you treat it as your replacement, you’ll find yourself replaced — not by AI, but by an engineer who used AI better. The advice for students: don’t use vibe coding to skip fundamentals. Use it to ship faster once you have them.

Three things he wants every young engineer to hear

In closing, Eslam refused to say anything new. He repeated what he’d already said — on purpose. Some things, he believes, are worth hearing twice.