ishanalone.in is a realized long dream.
I am a Lead iOS Engineer. I live in a world of compiled binaries and strict architectures. I generally avoid web development because the ecosystem feels like chaos.
But I always wanted this piece of online real estate.
This is the log of how I finally built it, and how LLMs and Agents helped me overcome the barrier that was stopping me.
The Barrier: The "Guilt" of Not Building
Hosting has always been affordable. Building was the hard part.
For years, I carried a specific kind of guilt. I am a product engineer. I’ve worked for product-based companies my whole career. I build complex mobile applications for a living. And yet, I couldn't build a simple website for myself.
It wasn't a lack of logic; it was a lack of bandwidth.
The Trap
I thought, "I'm an IT engineer, I should be able to do this easily."
The Reality
My actual job (iOS) demands rigorous, continuous learning. I physically could not spare the hours to master HTML/CSS or fight with WordPress configurations just to get a "Hello World" online.
I tried templates. I tried WordPress. But the fear of maintenance and the frustration of design always stopped me. LLMs and Agents turned that upside down.
The Stack: Irrelevant
People ask why I used Next.js. The truth? I didn't choose it.
I let the Agent (Google Antigravity) pick the stack. I don't care if it's React, Vue, or raw HTML. The Agent made the stack irrelevant. It handled the boilerplate, the Hostinger configs, and the Git autodeploy pipeline.
The Workflow
My process wasn't just "coding"—it was building.
1. The Manager (Setup & UI)
I hate CSS and configurations. So, I assumed the role of a Manager: "Make this button round. Fix the layout. Deploy it." The Agent did the heavy lifting. It allowed me to bypass the "Guilt of Not Building" and just ship the product.
2. The Senior Engineer (The Physics)
LLMs are great at syntax, but bad at physics. When we built the Hero section, the Agent used high-def animations that crushed the DOM on mobile data. Here, I assumed the role of the Engineer. I architected an array-based pre-loader to manage memory and DOM pressure.
The Lesson
A friend (an ML entrepreneur) once told me:
"The LLM is a magic wand. You can be Hermione, or you can be Ron."
If you don't know the fundamentals, you are Ron with a broken wand—expecting magic but getting chaos. Because I understood the engineering (the physics), I could guide the wand like Hermione.
The Philosophy: Mastery & Limits
Platforms are just tools. I have enough mastery in one tool (iOS) to reverse-engineer the syntax of another (Web).
However, I also know my limits. I do not know Service Architecture. Even if you gave me the most powerful tools in AWS, I would fail miserably because I don't think in microservices.
That is why I chose Hostinger. It allows me to stick to the Monolith—the architecture I master—while using Agents to handle the syntax I don't.
The Roadmap: RAG & LLM Engineering
I know this current setup won't fail, but I also know I can't stay in the monolith forever.
My Next Project:
- Local RAG-LLM on iOS: I will build a local Retrieval-Augmented Generation app. This keeps me in my comfort zone (iOS Monolith) while I learn the fundamentals of LLM engineering.
- Cloud Migration: Once I understand the logic, I will move it to a cloud-based RAG product. That is when I will force myself to learn Microservices.
The Utility: Forward Only
This site is the stage where I will document that transition.
I could easily populate this space with my past learnings—years of iOS notes and tricks. But that would be a waste of time. There is too much new stuff to learn. This site is not an archive of what I did; it is a log of what I am doing.
Scaling is a different argument for a different day. For now, the goal is ownership.