Jatinder S Luthra

Technical adventures in startups and enterprises.

I am human and I write about whatever I’m thinking hard about. Sometimes that’s technology. Sometimes it’s a book, or something I got badly wrong, or a conversation I’m still thinking about.

Anything another human can relate to.

Series

Sounds Familiar 1 part Complex ideas explained through everyday analogies Stack Confessions 5 parts What the infrastructure would tell you, if it could talk. The Art Of The Possible 1 part End to end builds of real workloads including working code, architecture decisions, and the reasoning behind every design call. The Third Option 6 parts A serialized story about a health insurer trying to build AI with secure data without betraying eight million patients. Every leader in the room is right. They are all still stuck.

Recent writing

Does your cloud bill feel like a fast food habit? Sep 15, 2026 4 min
Renting compute is cheap the first time and expensive the thousandth. A look at which workloads are worth cooking at home.
Build Your Own Object Storage with Oxide and MinIO Sep 9, 2026 7 min
Own your object storage with Oxide Cloud Computer and MinIO to meet your cost and sovereignty goals
When Network Talks Aug 29, 2026 4 min
Network discuss its life After AI
When Storage Talks Jul 13, 2026 5 min
Storage ponders on its life of being expensive, then cheap and now one of the asset
When Memory Talks Jun 20, 2026 5 min
Memory is no longer a commodity but leading the infrastructure decisions. Learn from memory itself
When Database Talks Jun 6, 2026 7 min
Database explains what it has been trying to tell you for years, and why nobody listens until something breaks.
When Data Talks May 25, 2026 6 min
Data has a confession to make, and a life story split into two eras: before AI, and after.
The Reckoning: Board Meeting Apr 11, 2026 7 min
Jessica asked the board for $1.2 million. She walked out with $300,000 and three months to prove she wasn't wrong.
The Dilemma of the Data Engineer Mar 27, 2026 5 min
Data Engineer explains how empty honey pot resolves the challenges of consistency, behavioral preservation, and realistic value libraries
The Dilemma of the ML Engineer Mar 13, 2026 8 min
ML Engineer, James Mitchell prediction models collapsed in production which were 89% accurate in development

All writing →