The thing that started it all
I was a teenager when I first encountered the idea that intelligence could exist inside a machine. Not simulated intelligence — actual emergent behaviour that even the people who built the system couldn't fully explain. The black box problem. The fact that the creators of some of the most powerful models in the world genuinely cannot tell you why the model made a particular decision — that baffled me. It still does. And that combination of bafflement and fascination is what pulled me headfirst into AI.
What sealed it was realising that mathematics — something I'd always loved for its precision and logic — could evolve into something as expressive and almost alive as a neural network. The structure of a neural network is genuinely beautiful to me. There's art in it. Not metaphorically — actually, visually, conceptually beautiful. I've always believed that anything done well has an architecture to it, a method, a form. And AI made that belief feel like more than just a personal quirk.
My goal, someday, is to build something like that from scratch. Something that surprises even me. For now, I take real pleasure in applying these models to real-world problems — building pipelines and workflows where I can rely on AI's intelligence as a genuine collaborator, not just a tool.
How I actually work
I am a data-first thinker.
Before I write a single line of code, I figure out what my data looks like. Input, output, and every meaningful shape in between. I think of them as lego bricks — discrete, well-defined data structures that I can reason about, pass around, and connect. Once I know my data, the rest of the architecture tends to follow naturally.
Even when I'm using AI agents to help plan or generate code, I drive the data design myself. I want to know exactly what a payload looks like before it leaves a service, and exactly what the receiving end expects when it arrives. I sketch these out explicitly — sometimes literally, sometimes just mentally — before building starts. This approach means I rarely get surprised by integration bugs or mismatched contracts halfway through a project.
I always build from scratch. Not because I'm opposed to frameworks or libraries — I use them constantly — but because I want to understand the bones of what I'm building before I layer anything on top.
Why the range
My portfolio spans RAG chatbots, OCR pipelines, WhatsApp bots, voice microservices, design systems, revenue collection platforms, and kiosk UIs. That wasn't a deliberate strategy. It's just how curiosity works.
I got into the world of AI and discovered RAG. I thought it was a genuinely lovely architecture — retrieval and generation working together as one coherent system. So I built one. Then I came across OCR and thought the same thing, so I built that too. WhatsApp bots taught me something different: you can build a real product without any traditional UI at all, just a conversational interface over an existing messaging platform. That idea excited me.
Then I started thinking about how to make RAG even more efficient — what if you removed the database layer entirely, and just crawled the website live every time? That became LiveMind. Each project was less of a planned decision and more of a question I wanted to answer by building the answer.
The art-and-science thing
I've spent most of my life refusing to choose between the two, and I've come to believe the separation is mostly artificial anyway.
Science done well has aesthetic qualities. A clean system architecture is satisfying in the same way a well-composed painting is satisfying. A well-designed API feels elegant. Code that reads like prose is genuinely pleasurable to write. I was the lead singer in my school choir. I spent years sketching and painting as a child. I follow modern artists on social media and think about visual composition the way I think about software architecture — where does the eye go, what's the flow, what does the structure communicate?
To me, building something — anything — is a creative act. You're making something that didn't exist before. Whether it's a model, a microservice, or a melody, the fundamental experience is the same.
What I'm looking for
I do my best work on collaborative teams with good management — places where teammates are people you can genuinely rely on, not just people you technically work alongside. That distinction matters to me more than company size or industry. I want to work somewhere where people take the work seriously and take each other seriously.
In terms of what I want to build next: I'm deeply interested in agentic workflows — systems that can reason, plan, and take action across multiple steps without constant human steering. I've been using existing MCPs (Model Context Protocols) for a while and I want to start writing my own. There's something compelling about defining exactly what capabilities you give an agent and watching it use them in ways you didn't fully anticipate.
Outside of work
Music accompanies almost everything I do — working, reading, travelling, thinking. My taste spans languages, genres, and moods because I'm attracted to feeling, not category. K-pop is a genuine love, not an ironic one.
I'm a film and television enthusiast with specific taste: ambitious sci-fi (Interstellar, Tenet), intelligent character work (Sherlock), emotionally rich storytelling (When Life Gives You Tangerines, The Good Bad Mother). I also watch anime, the occasional German show, and can go deep on K-dramas. My reading covers science fiction, young adult fiction, and — I'll be honest — fanfiction. I've been reading and writing fanfic since my teenage years and I maintain a Tumblr account for it that I will absolutely not be sharing.
I keep a small, close circle of friends who I value deeply. And I love road trips — I grew up travelling by road from Pune to Hyderabad, Hyderabad to Nashik, and further, which built a habit I've never dropped. I've done Goa, Pondicherry, Chandigarh, and various cities across Maharashtra. The journey matters as much as the destination.
At the end of it, I'm an engineer who thinks like an artist, a learner who builds to understand, and someone who finds genuine excitement in problems that don't have obvious answers yet.