The wrong person to tell you about my brother
ยท 12 min read
Let me get the disclosure out of the way, because everything below is worthless without it. Raiyan Siddique is my brother. He is the CTO and co-founder of my company. I am, by any reasonable standard, the least trustworthy person alive to assess him.
So do not take my word for anything. Every claim in this post is checkable. He has a page listing what he has built, a LinkedIn, and code running in hospitals. Go and look, then decide.
What I want to write about is not whether he is good. It is the question people actually ask me, which is how somebody gets here at eighteen.
Genius is the wrong word
People reach for it because it is quick, and because it lets them stop thinking. If he is a genius then he is a different kind of creature and nothing about his path applies to anyone else. That is a comfortable conclusion and I do not think it is true.
Here is the boring arithmetic instead. He started writing code at eleven. He is eighteen. That is seven years. A computer science graduate in India typically starts at eighteen and finishes at twenty two, which is four years, most of it spent on assignments nobody deploys.
He has been compounding for seven years, and for the last two of those he has been shipping software that clinicians use on patients. Measured in hours actually spent making things work, he is not ahead of his age. He is on schedule for somebody who has been doing this for seven years. The surprise is only in the calendar.
Calling it genius explains nothing. Calling it seven years explains almost everything.
The thing that mattered more than talent
I have watched this closely, and if I had to name the single biggest factor it would not be intelligence. It is that nobody ever successfully told him he was not qualified yet.
In India that is unusual, and it is the real bottleneck for most people. The system runs on credential order. You finish the degree, then you apply, then somebody decides whether you are allowed to touch anything real. By the time permission arrives you have lost your teenage years to preparation for a gate.
Raiyan skipped the queue by building things that already worked and letting the work do the arguing. Before Zero Dementia he built hearAI, a hearing screening system with custom hardware, for a neurosurgeon at a government hospital. Nobody hired an eighteen year old to do that. He turned up with something functioning.
Which, and I recognise the irony of writing this, is the same move I used to get into research labs. Neither of us found a way through the front door. We both worked out that a finished thing is a better argument than a CV. I did not teach him that. He watched it work and drew his own conclusion, which is a different and probably more durable kind of learning.
He left college
This is the part people react to most, so I want to give his reasoning properly before I argue with it.
He left because he decided there was nothing there to learn. Not that it was hard, not that he was failing. That the syllabus was years behind what he was already doing every day, and that sitting through it was the most expensive thing in his week.
I think the honest version of his case is stronger than people assume. Engineering curricula in India get revised slowly, and the gap between a syllabus and current practice in machine learning is measured in years, not months. A student who is already training models and running production infrastructure is not being stretched by an introductory paper on material that was settled before he was born. He was being asked to spend his most productive hours proving he knew things he had already used in anger.
And the cost was not the tuition. It was the hours. Attendance requirements, lab submissions, examination cycles, all of it billed against the only resource that actually mattered to him, which was uninterrupted time to build. He looked at that trade and decided it was a bad one.
His objection was never that it was difficult. It was that the price was his time and the return was a certificate saying he knew things he had already shipped.
Why I still would not repeat that advice
I finished mine. B.Tech in biotechnology, all four years, and I do not regret it. So the two of us are a natural experiment with a sample size of two and no control group, and you should weigh my view here knowing I chose the other path.
Three things make me cautious about turning his decision into a general rule.
You are only hearing from the ones it worked for. Dropping out is loud when it succeeds and silent when it does not. For every person whose exit looks visionary in hindsight there are many who left for the same reasons, with the same confidence, and simply lost several years. Their write ups do not exist. Any inference you draw from stories like his is drawn from a sample that has already been filtered.
A degree is an option, not just a signal. This is the argument I actually put to him. The certificate is largely useless as proof of skill, agreed. But it is a key to a set of doors that are hard to open otherwise, including graduate study, some visa categories, and employers whose filters are not negotiable. Options have value even when you never exercise them, and he gave up a real one. He knew that and accepted it, which is different from not having thought about it.
Boring is not the same as useless. The material he found tedious was largely foundational. Algorithms, theory, the mathematics underneath the libraries. Those are precisely the areas where somebody who learned by building ends up with holes, and the holes do not announce themselves. They show up later as a subtly wrong approach to a problem that a formal grounding would have made obvious. I expect him to hit some of those, and to have to close them the slow way.
The thing that actually made his decision defensible is one that gets left out whenever the story is retold. He had traction first. He was already building systems people depended on before he left, so he was walking away from one path while standing on another. Leaving because the syllabus is dull, with nothing built yet and a plan to start afterwards, is a completely different bet wearing the same clothes.
What he actually built, so you can judge for yourself
The whole Zero Dementia technology stack is his. A Flutter capture app that records voice, eye movement, facial micro-expressions and cognitive task responses at controlled sample rates. A signal pipeline using openSMILE, MediaPipe and librosa that pulls 187 biomarkers out of a single session. A Node.js and Postgres backend orchestrated with Redis, Celery and Airflow. The training and inference infrastructure on SageMaker.
Outside that, PRESAGE, which predicts cardiac deterioration from wearable heart rate signals using conformal prediction, validated against MIMIC-IV. An AI native ERP exposing 389 tools over MCP. A weather intelligence API for an IIT Delhi incubated startup. Around thirty production systems in total, four papers under review, and co-inventor on the multimodal biomarker fusion patent.
I am listing these because they are artefacts, not adjectives. You can look at a running system and form your own view. You cannot do that with the word genius.
The unfair advantages, stated plainly
Any honest version of this has to include the parts that were not merit.
He had an older brother already inside health research who could open clinical doors, make introductions, and hand him problems that mattered. When he built hearAI, the neurosurgeon was reachable partly because of relationships I already had. He had a family that treated a kid disappearing into a laptop as acceptable rather than a problem. He had electricity, a machine, and no obligation to earn money at fifteen.
Plenty of people with more raw ability than him do not get any of that. I am not going to pretend the playing field was level, because pretending that is how these stories become useless to the people reading them.
What he is like to work with
He is my CTO, which means I get the version of him that is not on any profile page. It is not uniformly easy.
He is faster than anybody I have worked with, and the speed comes with a preference I have learned to argue about. Given a system that is not behaving, his instinct is to rebuild it rather than repair it, because rebuilding is more interesting and he is quick enough that it often is genuinely cheaper. Sometimes that is the correct call. Sometimes it means a component that was merely inelegant gets replaced by one that is new, and new means untested.
He under documents. Code that is obvious to the person holding the whole design in their head is not obvious to anybody else, and for a while the whole design existed in exactly one place. We have had to make that a rule rather than a request, because a stack that only one person can maintain is a risk to the company regardless of how good that person is.
He has very little patience for process, which is the same trait that got him out of a lecture hall and into production, pointed at me instead. Some of that friction is useful. Process that cannot justify itself should be challenged. Some of it is just a cost of working with somebody who has never had to persuade a committee of anything.
I am including this because a profile without any of it is marketing. The strengths and the friction are the same trait viewed from two sides, and if I only described one side you would be right to discount the whole piece.
What I genuinely cannot tell you
I am not the right reviewer for his code. I know enough to read a pipeline and follow an architecture, but I have never been the senior engineer who tears it apart properly, and he has never had one. That is a real gap and it is the thing I would fix first if I could.
Starting at eleven also means there are foundations he learned sideways rather than formally. Somewhere in there are gaps that will surface later, in the way they always do for people who learned by building rather than by curriculum. He will find them the hard way.
And I have no idea whether this pace is sustainable, or what it costs him. Anyone who tells you they can assess that in their own sibling is not paying attention.
What I would actually take from this
If you are looking for the transferable part, it is not the age. It is that the gate is usually softer than it looks, and the way through it is almost always a working thing rather than a request.
And if you have a fourteen year old in the house who has gone quiet and is building something you do not understand, the most useful thing you can do is nothing. Do not redirect them toward exam preparation. That is the whole intervention.
He is at linkedin.com/in/raiyanz if you want to check any of this, or hire him, which I would rather you did not.