Zen Mascot

I build simple solutions
for complex problems.

Hi, I'm Ghazi. I build lightweight tools for developers, focused on performance, clarity, and privacy-first systems.

See my projects
file_type_python file_type_mysql
Currently Building

docmd

Build production-ready documentation from Markdown in seconds. No React, no bloat, just content.

DocumentationMarkdownCLIBunNode.jsTypeScript
Check it out →
docmd

memoryblock

Private AI workspaces that run locally, using fewer tokens and less compute.

AILLMPrivateLocalWeb AppCLILibraryBunTypeScript

AIMonitor

Native macOS menu bar app for monitoring AI usage, quotas, and provider status in real time.

macOSSwiftAIMenu BarOpen Source

Latest Rollups

View all →
Sep 24, 2026

Speculative Decoding Hits Vision Models: What Developers Need to Know

Liquid AI's DSpark draft model accelerates vision-language inference 2-3x on edge devices and 20x on H100s. Here's what it means for your stack.

Sep 24, 2026

Building Fast Diff Surfaces: How GitHub Handles Million-Line Pull Requests

Inside the architecture behind rendering massive pull requests with hundreds of comments performantly. A deep dive into virtualization, geometry, and engineering trade-offs.

Sep 24, 2026

GPT-6 Luna is ridiculously cheap, and the AI price war just got real

OpenAI and Anthropic dropped massive price cuts today. Here's what it means for developers building with LLMs in 2026.

Sep 24, 2026

Scaling MuJoCo to 2048 Parallel Simulations on GPU

How MJWarp bridges MuJoCo and NVIDIA Warp to parallelize robot simulations for reinforcement learning at scale, without rewriting physics code.

Sep 23, 2026

Why Slack Channels Are Becoming Developer Environments

Slack's Code Channels feature breaks down silos between developers and AI agents, merging code writing and review into a single collaborative space.

Sep 23, 2026

Making AI Evaluations Reproducible: AISI and EvalEval's Open Infrastructure

How AISI and EvalEval are standardizing AI evaluation reporting through shared schemas and open platforms to improve reproducibility and research reliability.