About
Two years ago I was on the phone with my mum, spiraling. I couldn't decide which branch of CS to commit to. The field felt too broad, and I kept getting tossed around like the wind: ML one week, data science the next, backend after that, then nothing but LeetCode, then wherever I happened to get hired. I never really held my own in anything.
Then, one night shift at my internship, head wonky from fighting sleep, a client came by to drop off devices for certification. He asked what I studied. CS, I said. A few questions later we were standing at the PXA analyzer, and he told me the people who can program these things are extremely valuable: custom operating systems, recurring servicing revenue. I didn't know what bare metal was yet, and I won't pretend that was the moment it all clicked; if it were, I'd have walked away with a clear direction, and I didn't. But it stuck. Building the underlying infrastructure, writing the algorithm that powers the thing. I wanted that.
Now I work across infrastructure and reliability. At Red Hat's Konflux I built a reproducible-build pipeline so tampered builds can't ship — the same class of supply-chain attack behind SolarWinds. I did an SRE rotation under Meta engineers, learning to read a live system with nothing but grep, ps, and journald. And I've built projects like a C++ percentile engine that holds a thousand requests a second under 300ms. I care about the boring, load-bearing layer because when it's done well nobody notices, and when it isn't, everyone does.
Lately I've been pulled further down: toward embedded and RTOS, and, further out, quantum. I don't fully know where that leads yet. But I've never been able to stop asking what the layer beneath the one I just learned actually does. I'd rather build the floor everything stands on than the thing standing on it.