Most AI writing covers what’s possible. Fieldwork covers what actually happens.
I’m Mayank Bohra, a backend engineer in Delhi building Rehearsal (a voice AI interview platform, 2,000+ users) and Highlyt (an AI-assisted reading tool). My work has been published in Towards AI. I write about what I learn while building. Not in hindsight, not from memory. While it’s happening.
What you get here
Build logs from real production systems. Post-mortems on architecture decisions that didn’t hold up. How I use Claude, MCP, and context engineering to ship AI products without breaking everything else. The gaps between what AI tools promise and what they actually deliver. The occasional piece on how to think clearly in a period where thinking is the one thing AI can’t do for you.
Who this is for
Engineers and builders tired of AI content that reads like it was written by someone who has never deployed anything. If you’re shipping real AI products and want honest notes from the field, this is for you.
What to expect
One post a week. Practical, specific, written to be useful to someone sitting where I was six months ago.


