Ismam Khan (@K_Ismam) header
Ismam Khan (@K_Ismam) profile picture

Founder @ Logic Gate Software Solutions Most businesses don’t need more tools — they need systems built for scale I make that shift happen

Dhaka, Bangladesh
Joined February 2017
Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)Photo by Ismam Khan (@K_Ismam)
Used to plan in big releases — weeks of work, one big launch. Started forcing the team to ship in much smaller pieces this year, sometimes daily. Felt slower at first. Wasn't — bugs got caught faster, feedback came in sooner, and nobody was sitting on three weeks of unreviewed work hoping it was right. Small and frequent beats big and occasional almost every time. Wish we'd done this years earlier.
7
Used to measure a good week by how many hours I personally spent writing code. That number is close to zero most weeks now, and it took a while to stop feeling guilty about that. The actual job changed. It's less about my hands on a keyboard and more about whether the team has what they need, whether the right projects are even on our plate, and whether Logic Gate is the kind of place good engineers want to stay. Different kind of useful. Took longer than I'd like to admit to accept that.
2
Working on a healthcare platform right now where doctors and patients connect for live consultations. Early on we assumed everyone using it would be comfortable booking things on a phone. Wrong assumption. A lot of the actual patients are in rural areas and aren't comfortable navigating an app at all. So we added a moderator role — someone who can be physically present and book and manage a consultation on a patient's behalf, syncing it into the same system as if the patient had done it themselves. That one feature wasn't in the original brief. It came from actually picturing who'd be using this thing, not just who asked for it.
4
A developer executes a spec well — give them a clear task, they'll build it. Valuable when you already know exactly what needs building. A systems partner does something different. They question the spec before executing it, ask whether the request even solves the underlying problem, and think about what happens to this decision a year from now. The businesses that get burned are usually the ones that needed a systems partner but hired a developer, then wonder why the "finished" software didn't fix their actual problem.
1
Looking back at the last few weeks, the strongest technical decisions we made all came out of actual disagreements in the team — not consensus, not me deciding alone, genuine back and forth where someone pushed back hard on an approach. I used to think a smooth-running team meant everyone agreeing quickly. Now I think it means people feel safe enough to disagree loudly when it matters. Trying to protect that, not smooth it over.
Every shortcut taken to ship something fast is a loan. Loans aren't inherently bad — sometimes speed matters more than perfection. But loans come with interest, and the interest on technical debt shows up as features that take longer to ship and bugs that multiply in unrelated places. The businesses that get hurt aren't the ones that took on technical debt. It's the ones that never tracked it and woke up one day unable to ship anything quickly anymore. If your team keeps saying "this should be quick" and it never is, that's not bad developers. That's unpaid interest on decisions made months ago.
2
We've been talking through an idea internally — a tool we'd call Growth OS, basically a structured system that takes a business idea and walks it through research, positioning, and an actual execution plan instead of just "here's a content calendar template." It's purely a discussion right now, not something we've started building. But the question we keep coming back to is whether something like this would have actually helped us grow Logic Gate faster in our own early days — and the honest answer is probably yes. Might build it. Might not. Wanted to think out loud about it here instead of pretending we have it all figured out.
3
"We've always done it this way." I hear this constantly, usually right before I find a process that's quietly been bleeding time or money for years. It's never said with malice. It's said with comfort. Familiar and efficient are not the same thing. Every time I hear that phrase, my next question is always: why, and is that reason even still true today? Half the time, the original reason disappeared years ago. The process just never got the memo.
We have a team that handles most client communication day to day. I still jump into a thread directly every so often, especially early in a relationship. Not because I don't trust the team — because there's something a founder's direct attention signals that a process can't replicate, especially for a client taking a real risk on a company our size. I'm trying to be more deliberate about when I do this instead of doing it out of habit or anxiety. Still figuring out the right balance.
1
Nomad started from a real moment — a student studying abroad in Barcelona who had no easy way to find people in her program or figure out what to actually do in the city. That's still the core of it: a social app built entirely around travel and exploration, where you post your journeys day by day, discover what's happening near you, and find people in the same program or city as you. We just started rolling out phase two — group trips where people plan together before they even leave, a visited-countries tracker that turns your travel history into something you can actually show off, and a much smarter feed underneath all of it.
1
16
Before spending money on new software, I run through three questions. One: does an existing tool already solve most of this? If yes, buy, don't build. Two: is this core to what makes the business different, or is it generic infrastructure? Generic — buy. Core to your edge — build. Three: will this still look the same in a year? If you're not sure, wait, and keep doing it manually a little longer. Most businesses skip straight to "let's build something custom" without running this filter and end up paying real money to solve a problem a $30/month tool already handles. We tell clients this regularly, even when it costs us the project. The trust it builds is worth more than the build would have been.
9
Sat in on a debugging session this week where two of our junior developers caught something our more senior person had missed for two days. Took them forty minutes. Fresh eyes catch things tenure misses. I used to think experience was just about seniority. It's also about not being too familiar with your own code to see what's actually wrong with it. Good reminder to keep rotating people onto things, even senior people onto junior-owned code occasionally.
4
Most dashboards don't create clarity. They just make confusion look organized.
Building an auction platform where thousands of people might bid on the same item at the same second, and it has to stay accurate down to the millisecond. The boring infrastructure decision — processing bids in strict order instead of letting them race the database — matters more than anything visible on screen.
2
Hiring our first two senior developers who could run projects without me in every conversation was uncomfortable before it was useful. Had to practice not being the answer to every question. It felt like losing control. But it was actually the first sign this could become bigger than just me.
9
A feature does one thing. A system decides how all the features talk to each other. Most software problems aren't broken features — they're five features that each worked fine alone, and nobody thought about what happens when they meet. Like an invoice feature and a payment feature that don't agree on what "paid" means.
1
Spent part of the weekend just using our own products like a regular person would, not testing them. Different bugs show up that way. A test suite tells you the code works. Actually using it tells you whether someone would enjoy it. Humbling, in a good way.
4
A client once wanted us to skip straight to building, no questions asked, start Monday. I said no — that's exactly how you build the wrong thing fast. We lost the project. Still think it was the right call. The clients who stay longest are the ones who let us ask annoying questions first.
5
Built Dokanify because small business owners we knew were drowning in DMs trying to sell through Facebook and WhatsApp with no real storefront. The idea: sign up, tell us what you sell, get a real store fast, cheap enough that it doesn't scare anyone off. Still running today, and it taught us more about building secure multi-tenant systems than almost any client project has.
5
One question before I say yes to any project: "What happens at 100x the users you have today?" Filters more than any contract clause.
🤖 Made with AI
1