Adnan Ahmad

Blog

Deep dives into AI engineering, distributed system design, and practical lessons from scaling software architectures.

Building MermaidAI: An AI Diagram Studio with Gemini 3.6

Explore how MermaidAI Studio pairs Next.js 15, Gemini 3.6 Flash, Monaco Editor, and Mermaid.js v11 to transform text into live system architecture diagrams.

I Built an AI Course Generator with Gemini & Next.js

An open-source AI course generator built with Next.js and Gemini. Covers the generation pipeline, offline-first IndexedDB sync, and dual Gemini/Ollama provider support.

Unlocking Native Computer Use: Setting up CUA with Gemini 3.5 Flash

Step-by-step guide to setting up the trycua CUA Driver with Gemini 3.5 Flash and Model Context Protocol (MCP) for native macOS desktop automation using the Antigravity IDE.

How AI & LLMs Are Rewriting Software Engineering

From Spec-Driven to Eval-Driven development, explore how AI and LLM agents are reshaping traditional methodologies like TDD, BDD, and DDD — and what it means to be a software engineer in 2026.

Building a Safe, Harness-Agnostic Text-to-SQL Engine

Why standard SQL generation fails in production and how to build a harness-agnostic, secure database translation library with automatic Entra ID token authentication.

Designing Agent-Native Systems: A Deep Dive into StudyPal

Exploring the two-layer plugin architecture, tool registries, and stateful capabilities behind StudyPal — an intelligent learning companion built on autonomous, goal-driven execution loops.