No computer at home. No coaching centre. No programming course.

Just curiosity, a search bar, and years of trying things until they worked.

Before any of this was possible

I didn't grow up knowing I'd build websites for a living.

Everything I know today, I learned largely on my own through YouTube, Google, documentation, experimentation, and a lot of trial and error.

Before I got into technology, I did whatever work was available.

I sold vegetables. Worked at a dairy. Then as a technician, a fitter, and a helper. I've handled water tanks, mixed cement, and spent time working in a perfume factory.

I sold pakodas and chole kulche and roasted chana from a cart. Worked at a grocery counter, then a pan shop. Looked after goats and buffaloes, cut grass for them, and did farm work whenever there was work to do.

None of it looks like a straight line to where I am today.

But those years taught me something that has stayed with me: keep working, keep learning, and keep looking for a way forward.

An education that kept starting over

School never followed a straight path for me.

I moved between government and private schools, eventually studying in a government school through class 8. Years later, I got an opportunity to join a private school for class 9. I attended for a few days before the lockdown disrupted everything.

I passed class 10 during that period. That was effectively the end of my formal education.

No computer science degree. No programming certificates. No conventional route into software development.

Almost everything I use in my work today came from learning outside the classroom.

The moment curiosity took over

It started with something simple: watching someone in my village use a computer.

I didn't understand what they were doing. I just knew I wanted to learn.

There was no teacher or structured learning path waiting for me. I had YouTube, Google, and the willingness to keep searching.

If I wanted to understand something, I looked it up. If something interested me, I tried it. When it broke, I searched for the reason and tried to fix it.

That cycle became my education.

Honestly, it still is.

The blog that didn't pay off at first

I started blogging in 2017.

Like many beginners, I thought that having a website online would eventually mean making money. But publishing posts didn't automatically bring visitors, and without traffic, AdSense had little reason to pay me.

For a while, the results simply weren't there.

Instead of giving up, I kept experimenting. I tried different approaches, learned more about websites, and looked for ways to turn the skills I was developing into something useful.

I didn't have a clear business plan. I learned by building, observing what happened, and trying again.

What selling websites taught me

One opportunity eventually became a business of its own: building and selling AdSense-approved websites.

Over the years, I've sold more than 200 websites.

That experience taught me how to build sites with clean structures, useful content, working navigation, and the technical foundations needed to operate properly. It also taught me about the practical side of online businesses: what buyers value, what makes a site usable, and what needs to work before a project is ready to change hands.

Selling websites gave me a way to turn my skills into income. But I eventually wanted something different.

I didn't want to spend all my time building things just to sell them. I wanted to create products I could continue improving, maintaining, and growing myself.

The first real payment

Eventually, my own websites started earning through AdSense.

That was a turning point. After spending years trying to understand how websites worked, I had begun to understand another part of the equation: how a website could attract an audience and generate revenue.

From there, I explored other advertising networks, sponsorships, and different monetization approaches.

There isn't just one way to make a website sustainable. Advertising, sponsorships, and selling a site can all play a role, depending on the project and its stage.

More importantly, running websites taught me to think beyond publishing pages. A website is a product with users, technical requirements, operating costs, and business decisions attached to it.

Building something of my own

As I gained experience, I started investing more time in projects I wanted to keep rather than build solely for resale.

One of them is MemesTube.

What began as a meme platform grew into a much more involved engineering project. Over time, I worked on search, recommendations, APIs, content management, media delivery, caching, performance, and automation.

I had to learn how different parts of a production application fit together: how to retrieve relevant results, organize data, avoid unnecessary work, deliver media efficiently, and improve the experience for actual users.

Some problems required new features. Others required rethinking how an existing system worked.

I learned by solving those problems one at a time.

A real application with real users teaches you differently. Every change has consequences, and eventually, every important problem needs a solution.

You can explore more of my work on the projects page.

Different projects, different lessons

Over the years, I've worked across different types of websites and applications, from publishing platforms to local service websites and multi-site systems.

Each project brings different constraints. One requires better search visibility. Another needs a more efficient database query. Another needs an API integration, a new automation workflow, or a simpler user experience.

Working across different projects has helped me understand how to approach problems rather than rely on a single familiar solution.

The goal isn't to use the newest technology everywhere. It's to understand the problem well enough to choose an approach that works.

Learning by needing to

I never followed a formal course in PHP, databases, or APIs.

I learned these technologies because projects demanded them.

A feature was missing, so I learned to build it.

Something was slow, so I investigated the bottleneck.

Search wasn't returning useful results, so I learned about indexing, ranking, and relevance.

Two systems needed to communicate, so I learned how to design and integrate APIs.

As my projects became more complex, I moved beyond individual pages and scripts into backend architecture, database design, caching, data processing, and production infrastructure.

It's not always the fastest way to learn. But it means that much of what I know is connected to a problem I've actually had to solve.

Learning how search engines work

Building a website is one challenge. Getting people to discover it is another.

Over time, I learned about search intent, content structure, internal linking, technical SEO, backlinks, indexing, and performance.

Much of that knowledge came from working on my own websites, studying their results, and experimenting with different approaches.

SEO is never completely settled. Search engines change, websites evolve, and something that works in one situation may not work in another.

That experience also taught me to look at websites as complete systems. Content, technical infrastructure, user experience, and distribution all affect how a product performs.

Where automation and AI fit in

As I worked across more websites and publishing workflows, repetitive tasks became harder to manage manually.

That pushed me toward automation.

I started connecting APIs, databases, scripts, and external services to reduce repetitive work and make my systems more efficient. Eventually, I began incorporating AI into those workflows.

I've built automated pipelines that gather source material, process content, generate supporting media, and prepare material for publication across websites and languages. I've also experimented with AI agents that coordinate multiple steps in a workflow.

One example is covered in How I Run 49 Websites With One AI Agent.

My work with AI extends beyond generating text. I'm interested in how language models can interact with structured data, APIs, databases, and existing software to complete useful tasks.

That includes working with LLM APIs, embeddings, semantic search, retrieval-augmented generation (RAG), and agent-based workflows.

What interests me most isn't AI as a chatbot. It's AI doing useful work inside a larger system.

From websites to software systems

For a long time, I thought of myself primarily as someone who built websites.

Today, my work increasingly involves the systems behind them: backend logic, database queries, search infrastructure, APIs, automation, and the integration of AI into existing applications.

I work with technologies including Python, PHP, JavaScript, MySQL, and PostgreSQL. I've also worked with vector search, caching, content pipelines, and cloud-based services.

I enjoy working end to end, from understanding a requirement and designing an approach to writing the code, integrating services, deploying the result, and fixing what goes wrong in production.

I don't claim to know everything. Technology moves too quickly for that. What I value is being able to investigate an unfamiliar problem, learn what I need, and turn that learning into something that works.

Still learning, still building

More than eight years after I started blogging, my learning process hasn't changed much.

I search. I read. I experiment. I make mistakes, investigate them, and try again.

The tools have changed. The problems have become more complex. But curiosity is still what gets me started.

I still have plenty to learn, and that's part of what keeps the work interesting.

Where I am now

Today, I'm an independent developer building and operating web applications, backend systems, automation pipelines, and AI-powered workflows.

I continue to work on my own projects, build practical utilities on the tools page, and take on opportunities involving development and technical services.

My journey from selling vegetables and doing farm work to building search infrastructure and AI-driven workflows was never a straight line. I didn't have a traditional route into technology, and I didn't follow a predefined plan.

There was a blog in 2017. More than 200 websites sold along the way. Production systems built through years of experimentation. And projects I continue to develop because I want to see how far they can go.

I built my own way into technology one problem, one search, and one fix at a time.

And I'm still building.