Writing
Writing on LLM Engineering
Mayank Bohra's writing on production LLM systems, RAG, agent architecture, context engineering, and AI product trade-offs.
Focus
My LLM engineering writing is about the part that breaks after the prototype: context quality, evaluation, observability, latency, user trust, and the operational details that decide whether an AI system survives real use.
Where to Read
The original articles live across Medium, Towards AI, FAUN.dev, and Substack. I keep the full pieces there and use this site as the stable home that connects the themes together.
Read the articles
Towards AI · 7 min read
The Agent Prototype Trap: 15 Things That Break When You Hit Production ↗Towards AI · 7 min read
Everyone's Building with LLMs Wrong. Here Are the 10 Myths Killing Your Projects. ↗Towards AI · 6 min read
Stop Chasing Perfect Prompts. Build Context Systems That Actually Scale. ↗Towards AI · 5 min read
The Myth of Effortless AI Automation (Or: Why "Boring" is Harder Than It Looks) ↗Towards AI · 7 min read
Beyond the Prompt: What Google's LLM Advice Doesn't Quite Tell You ↗Medium · 6 min read
The Agentic Ops Headache: When 'Rollback' Means Complex Compensation ↗Medium · 3 min read
Clever Code, Constrained Power: What the DeepSeek Debate Reveals About Global AI ↗