I'm Varun Sharma. I spend my time at the seam between distributed systems and applied AI — the place where a clever model meets the unglamorous plumbing that decides whether it ships or stalls. Most of what I've learned came from operating things in production, watching them fail in interesting ways, and writing down the fix so I wouldn't relearn it the expensive way.
This site is where that writing lives. The pieces fall into three rough buckets: Systems, on the boring infrastructure decisions that compound; AI, on building and operating agent fleets without setting money on fire; and Engineering, on the habits and workflows that let a small team move fast without breaking the things that matter.
I'm partial to static sites, sharp interfaces, and code that a tired person can read at 2am. If something here is wrong or you want to argue about it, the fastest way to reach me is LinkedIn or GitHub.
What I write about
- Systems — durability, failure containment, and why static beats dynamic more often than people admit.
- AI — agent orchestration, inference economics, and the coordination problems that show up the moment you spawn more than one.
- Engineering — velocity, review loops, and shipping discipline.
Recent writing
- Why static beats dynamic for 99% of sites
- Building an Agent Fleet at Scale
- The economics of LLM inference at scale
- Shipping 50 PRs in a Week: What Actually Changed
You can subscribe to everything via RSS. No newsletter, no popups, no tracking beyond what a static host gives you for free.