About
I ran a companyI could not build.That was the training.
I am Izhar Ahmad Danish, an AI product consultant in Islamabad. I co-founded Triisum in 2021 and still run it. I cannot write production code. Six products have shipped anyway, and this is the long version of how that happened.
- NIC Islamabad · Cohort 12
- Co-founder & CEO, Triisum
- 6 products shipped
- $1M seed round supported
Triisum · February 2021 to now
The four years that look like a gap on paper.
Triisum is a travel eco-marketplace. I co-founded it in February 2021 and have been its CEO since. The work was the operator half of a startup and most of it was mine to do: 50+ customer interviews, 10+ white-label agreements, two OTA partnerships, two fintech integrations, and a team of 8+ across engineering, QA and operations. We were selected for NIC Islamabad Cohort 12 from a field of 1,110+ applicants.
None of that involved me building the product. I could describe it precisely and I could not make any of it exist. Every idea became a ticket, every ticket joined a queue, and the queue was other people who already had a full week.
What you learn in that position is narrow, and it is more useful than it looks. You learn to write the thing down properly, because the written brief is the only part of the process you actually control. You learn that a two-week estimate becoming six is usually not an engineering failure: it is the point at which two people discover they had been describing different products. And you learn to go looking for that gap earlier, in a document, where it costs an afternoon instead of a sprint.
I want to be careful about what that is worth on its own, because it is easy to tell this story backwards. Four years of writing specifications did not make me able to build anything. It made the change that came later possible. That is a smaller claim and it is the true one.
Propellus.ai · June 2025 to June 2026
A year on the other side of the table.
For a year I was Business Growth and Strategy Manager at Propellus.ai, working remotely with a Singapore team. I supported a $1M seed round and raised $150k+ of it. I built and ran a 30+ investor pipeline, brought in 80+ qualified leads, left 5+ deals in negotiation, and secured a letter of intent worth nine figures annually on a ten-year term.
That last number needs its caveat attached, so here it is before anyone else supplies it. A letter of intent is not a contract and it is not revenue. It is a signed statement that two organisations intend to do something, and the deal it describes still has to be closed by people who are not me. I list it because it is the largest thing I have put my name to, not because it is money in a bank.
The part of that year that matters to this page is much smaller. I trained 15+ people on that team to use AI tools, Claude among them. That was the first time I had to teach the method rather than simply use it, and teaching it is what turned a habit into a procedure.
[PLACEHOLDER: what training the Propellus team taught you about your own method that you had not noticed while working alone (one or two sentences in your own words)]
What changed
Building software became a specification problem.
And specification is an operator skill. That is the whole thesis, and it is the one claim everything else on this site rests on.
When the AI coding tools arrived I used them the way most non-technical people do (a sentence at a time, like a search engine) and I got back exactly what that earns. Something that runs, looks roughly right, and is wrong in a way that takes longer to unpick than to rewrite. I decided the tools were built for engineers and went back to the queue.
[PLACEHOLDER: the specific moment you realised the brief was the lever: what you were working on, what you tried differently, and what came back (three or four sentences in your own words)]
What I do now is write the brief I had already spent four years learning to write for humans. The constraints. The edge cases. The file it goes in. What done means. What must not be touched. The model executes that literally, and literal execution is the point: it exposes every gap in the specification the same afternoon you wrote it, rather than at a demo six weeks later.
This is not what "vibe coding" means, and I would rather lose the search traffic than blur it. Andrej Karpathy coined the term on 2 February 2025: "There's a new kind of coding I call 'vibe coding', where you fully give in to the vibes, embrace exponentials, and forget that the code even exists." As he described it, you accept the diffs without reading them, paste the errors straight back in, and let the codebase grow on its own. What I teach is the opposite discipline. The specification is the work, and the distinction between those two things is most of the value; the long version of that argument has a page of its own.
What I do now
Four things. Two of them are free.
Consulting is the paid work and it is the smallest part of the week. Most people should start with the free half and never book anything.
01
Consulting
Founders and product teams
Three engagements at three depths: a free 30-minute strategy call, a two-week Build Sprint on your real codebase, and a 90-day programme for teams where the licences got bought and nothing changed. All three start with the same free call, and I will tell you on it which one you need, including when the answer is none of them.
See the three engagements02
Courses
Self-paced
The structured version of the same method, for people who would rather learn it than hire it: Claude Code for founders, AI product management, and Claude for marketers.
See the courses03
The vibe coding page
Free · no signup
Free, and the honest version: where the term came from, what it actually described, what I teach instead, and why the difference is the entire point. If you only read one thing here, read that one.
Read it04
The newsletter
Weekly · free
One email a week on what is actually working, written out of real client work rather than out of a feed. Free, and you can leave in one click.
Subscribe
Credentials
The résumé, stated plainly.
Every line below is checkable from my LinkedIn, my GitHub, or a product you can open in a browser.
What I have actually done
- Co-founder and CEO of Triisum since February 2021: a travel eco-marketplace, and a team of 8+ across engineering, QA and operations.
- Selected for NIC Islamabad Cohort 12 from 1,110+ applicants.
- Six products shipped: LittleOS, Liquid Finance, Safar, ThumbSnap AI, an autonomous publishing pipeline, and the Triisum Eco-Ambassadors app.
- Business Growth and Strategy Manager at Propellus.ai, June 2025 to June 2026: a $1M seed round supported, $150k+ of it raised.
- 4+ years in high-growth startups, five consulting clients advised, and 15+ people trained on AI tools.
What is not on that list: a computer science degree, an engineering job, and a line of production code I wrote myself. My GitHub bio is the accurate short version: "I'm non-technical, learning and teaching non-technicals vibe coding and automations".
The numbers
Four figures, all of them checkable.
Nothing here is modelled, projected or rounded up from something adjacent. If a number cannot be verified from my GitHub or my LinkedIn, it is not on this page, which is also why there is no revenue figure and no student count.
- Products shipped
- 6
- Clients advised
- 5
- People trained on AI tools
- 15+
- Following on LinkedIn
- 9K+
How I work
Islamabad, and remote everywhere else.
I am based in Islamabad, Pakistan, and I work with clients remotely, worldwide. Pakistan Standard Time is UTC+5, which puts most of a European working day inside mine and moves the US East Coast into my evening.
[PLACEHOLDER: the hours you will actually hold for US-timezone calls, and whether engagements run async with live sessions or the other way round (two or three sentences)]
The first call is 30 minutes and it is free. You describe what you are trying to ship and where it is stuck, and I tell you what I would do about it. There is no deck and there is no pitch. If I am not the right fit I will say so in the first ten minutes, and most of these calls end with somebody getting a useful answer and not spending any money.
Two things I will not do. I will not take an engagement where the honest answer is that you need engineering capacity rather than a method, and I will not put a delivery date on a scope nobody has written down yet. Both of those are how this work goes wrong, and both are avoidable in the first conversation.
01
Before: one paragraph from you
What you are building, where it is stuck, and what you have already tried. One paragraph is genuinely enough. If there is a repository or a live URL, send that as well and I will have looked at it before we speak.
02
During: thirty minutes, no deck
Most of it is me asking questions. The rest is a direct read on whether your bottleneck is specification, scope, or genuinely engineering capacity, and the one workflow I would change this week if it were mine.
03
After: a recording and a written summary
You get both, whether or not we go any further. If the answer is that you do not need me, the summary says that in writing, and you are free to go and act on it without me.
Next
Book the call, or read for a few months first.
Thirty minutes, free, and no obligation at the end of it. If you would rather not talk to anybody yet, that is a reasonable position: the newsletter and the six products are the whole of my evidence, so go and look at those instead and decide in your own time.
P.S. If you are non-technical and you have been told that is the reason you cannot ship, read the vibe coding page before you book anything. It is free, it is the argument in full, and it may be all you actually need from me.