“What do you actually do?” People have asked me this for years, my relations most of all, and my answer never quite lands, no matter how I phrase it. The trouble is not that I lack an answer. It is that the honest one sounds absurd: I am an engineer who refused, on principle, to become any particular kind of engineer.

That refusal was deliberate. To specialise seemed to me a small tragedy, like marrying young and forever wondering about all the marvellous problems you never courted. So I turned down the corporate label, and the pretentious posture that tends to come stapled to it, and gave myself instead to the one thing I genuinely enjoyed: solving puzzles.

The first puzzle found me at university, where I tumbled by happy accident into image processing. I read the papers. I redid the mathematics by hand¹ and then helped developers at a string of startups turn that arithmetic into something that actually ran, in Python and in C++. They were splendid years, and they left me with one article of faith I came to hold rather fanatically: that anything a human mind had dreamt up, my mind could learn. Nothing lay beyond the reach of learning.

That certainty was promptly put to the test. Fortified by it, I strode into a blockchain department — and was fired by the end of the week.

I declined to take the hint; one cannot, after all, fire a belief. So in 2021 I joined Nord Finance and spent three months learning to build, test, break, and deploy. They offered me a full-time post, a flattering gesture, but my freelancing had by then grown plump enough to feed itself, and I let it.

That same year brought the BankX contract, and I have scarcely looked over my shoulder since. Its mathematics, its tokenomics, its architecture were all gratifyingly complicated, more than enough to occupy a restless mind for several years.

But restless minds outlast their puzzles. By 2025 the major work had drawn to a close; we had tested and deployed very nearly everything. So, after a break² I joined a family friend to build Ctrlbit, an SDWAN product for Mikrotik hardware here in India. It let me learn the magic behind the internet and bend it to my own ends, which was a quiet thrill.

All of which brings me, at last, to the point of this page. Solving a hard problem means outwitting it. Explaining it means something stranger and better: you give the answer away, and somehow you keep it too. That second thing is rarer, harder, and the only kind of work that has ever held me. It is, in the end, what I actually do.