About
Saksham Goyal
I like building things.
Not just writing the code behind them, but taking an idea that is still a little vague, figuring out what it should become, and turning it into something real.
I've always been drawn to the messy part of building, the part where there isn't a clear answer yet. What should we build? What can actually be automated? Where does AI genuinely help? What's the simplest system that can make it work?
I started in AI research, then built the backend to serve the models, learned cloud to deploy them, and picked up the rest along the way. Mostly because that is what building a product takes. These days I work across AI, backend systems, and product engineering.
I enjoy going deep technically, but I'm equally interested in the problem on the other side of the code.
How I work
The why before the how.
I tend to start with a problem, break it down, and work backwards toward the simplest useful solution. Sometimes that's an LLM. Sometimes it's a boring database query. Sometimes it's a completely different approach.
I'm not particularly interested in using AI for the sake of using AI. What excites me is finding the places where it actually changes how a product works, making something faster, more useful, or simply possible when it wasn't before.
I'm also comfortable figuring things out as I go. Startups have taught me that the job rarely stays inside a neat description, and I've come to enjoy that.
Working with AI
Faster, but still mine.
I use Claude Code, Cursor, and Codex most days. They get an idea into working software much faster than I could alone.
I wouldn't call it vibe coding, though. The logic and the architecture stay with me. I like knowing why a system is shaped the way it is, so the agents speed up the writing while I keep the design in check.
Interests
What I find interesting.
- AI agents that can actually get things done
- Systems that combine deterministic software with probabilistic models
- Making AI products reliable enough to trust
- Building the infrastructure underneath intelligent applications
- Turning complex workflows into simple products
- Finding the smallest possible product that solves a real problem
There's still a lot we haven't figured out. That's the interesting part.
Beyond building
Away from the screen.
I cook when I get the time (which usually means a recipe open on the counter, or a call to Maa).
I'll play just about any sport going, though badminton has a special place. Video games when there isn't time for that.
I travel when there is a reason to. Beaches, mountain treks, or somewhere with nothing much to do at all, usually with a spiritual stop along the way. The photo further up is one of those.
And a good amount of it goes to family and friends, which is less a hobby than a default.
The rest of the time I'm following some curiosity or other. An afternoon on a product idea, a weekend on some new technology, or far too long trying to understand how something works when I could have just used it.
I like startups, good products, thoughtful engineering, and people who are genuinely excited about what they're building.
Most of all, I like building things.
Elsewhere