CTO, Asia Pacific Japan & Greater China — Dell Technologies
Singapore · I write about where enterprise technology meets its consequences — and I build to test the argument.

Writing · LinkedIn · X · vExpose Blog

Most technology strategy is written for the moment it's adopted. I'm more interested in the moment it's tested — when the pilot has to reach production, when the architecture has to survive the outage, when the compute bill lands on the board's desk. That's where technology stops being a roadmap and becomes a consequence.

I write about three things: organizational resilience, and what a company discovers is actually load-bearing only in a crisis; the economics of compute, where efficiency is quietly overtaking raw capability as the real constraint; and technology sovereignty as a category rather than a slogan. Across the whole enterprise stack — infrastructure, data, and AI — not one fashionable layer of it. The leadership side lives at deepakwaghmare.com; the deep technical side lives at vExpose.

Two books are in progress along those lines: Who Controls Your Intelligence? — for CIOs and CDOs deciding who really controls their AI stack — and The Load-Bearing Core, for boards on what cannot be allowed to fail.

This GitHub is the workbench behind the writing — benchmarks, reference builds, and the code I use to check an argument before I make it in public. I've never trusted a claim I couldn't run.

• reversibility-classes — if your AI agent gets this wrong, can you undo it? An open spec (R0–R4) for how recoverable an agent's tool call is, with a registry of 536 tools across 20 MCP servers, each checked against its source. Disagreements welcome; they're how the spec gets sharper.
• vexpose-labs — the working code behind vExpose: benchmarks, reference architectures, and the scripts I use to check my own numbers before publishing.
• inference-economics — a transparent calculator for the real cost of serving LLM tokens: tokens per watt, dollars per million, and how far cloud prices sit above the energy floor.
• Infrastructure-as-code reference builds — Kubernetes, OpenShift on bare metal, and storage automation from years spent closer to the metal. Kept public because someone always needs a working example.

AI in production · AI agent governance · The economics of compute · Technology sovereignty · Cyber-resilience · Enterprise infrastructure & storage · Data platforms · Kubernetes / OpenShift

Most of my public energy goes to the ideas, not the repos — keynotes and panels across the region, a newsletter for CIOs and boards, and the two books above. The leadership writing lives at deepakwaghmare.com; the technical deep dives at vexpose.blog.

If you're working where technology meets the business — the economics, the resilience, the part that has to hold when everything runs on it — that's the conversation I'm always up for.