React · Next.js · TypeScript · Design Systems · Developer Tooling · AI-Assisted Engineering
I build production frontend systems, reusable UI infrastructure, developer tools, and engineering workflows that make software easier to scale, verify, and maintain.
My work has grown from shipping frontend features into frontend platform engineering: architecture, design systems, quality automation, reusable tooling, and AI-assisted development workflows.
I currently focus on four connected areas:
| Area | What I work on |
|---|---|
| Frontend Architecture | Scalable React/Next.js applications, migrations, data-flow boundaries, performance, maintainability |
| Design Systems | Component contracts, design tokens, Storybook, accessibility, visual regression, reusable UI infrastructure |
| Developer Tooling | npm packages, CLIs, deterministic automation, documentation and repository health tooling |
| AI Engineering Workflows | Agent skills, MCP-oriented workflows, structured delegation, verification and quality gates |
The direction is intentional: write code → understand systems → build reusable systems → automate verification → increase engineering leverage.
docs-healthcheck — ✅ Published
An evidence-backed documentation quality gate and deterministic repair CLI for Markdown repositories.
What it demonstrates
- Install-and-run CLI workflow with
npx docs-healthcheck - Deterministic Markdown analysis with explicit ERROR / WARN / INFO severity
- Central rule catalog with documented rationale, authority class, references and configurable severity
- Broken anchors, missing local files, heading structure, duplicate headings and repository-health checks
- Safe/confirm/manual repair policy with dry-run support and rollback/revert capability
- Realistic healthy, partially broken and severely broken Markdown scenario coverage
- Automated build, typecheck, tests, package validation and npm distribution
- Real-world validation against a repository containing dozens of Markdown files and more than a thousand links
- GitHub Pages documentation and browser playground
Links: Repository · npm · Live Demo
persian-date-native — ✅ Published
A TypeScript-first Persian (Jalali / Shamsi) date engine designed as a real distributable library rather than a demo project.
What it demonstrates
- Zero required runtime dependencies with optional Day.js integration
- Native JavaScript
Dateinteroperability - ESM, CommonJS, browser-global and CDN consumption targets
- Automated tests, build verification, package checks and multi-runtime validation
- GitHub Pages documentation and interactive browser tools
- Automated npm release flow with registry, unpkg and jsDelivr verification
- Reproducible performance benchmarking
Links: Repository · npm · Live Demo · Releases
My current public-project loop is: build → verify → document → publish → use on real systems → maintain.
|
Status: Active A vendor-neutral collection of AI coding-agent skills covering frontend engineering, Figma analysis, pixel validation, design systems, performance, testing, architecture reviews, monorepo migration, and UI/UX workflows. This project represents my work on turning engineering knowledge into repeatable agent workflows rather than one-off prompts. |
Status: Active tooling Developer workflow tools for selecting project context, generating changelogs, running local-model chat, and experimenting with local agent/MCP workflows. This work explores developer ergonomics, context management, local AI, and deterministic fallbacks. |
|
Status: Technical research A practical exploration of React internals through a simplified implementation of rendering, state, effects, and element creation. The goal is not to reproduce React, but to understand the mechanics behind the abstractions used in production frontend work. |
Status: Earlier OSS A Markdown table-of-contents generator and one of my earlier public developer-tool projects. Its evolution into |
Some of my strongest engineering work lives in private/company codebases, so I describe the engineering problems rather than exposing proprietary implementation details.
- Design-system engineering: shared component architecture, public component contracts, tokens, Storybook quality, accessibility, visual regression and responsive validation
- Large React/Next.js applications: architecture audits, migrations, dependency cleanup, performance analysis and maintainability work
- Frontend quality systems: CI/CD checks, bundle and runtime analysis, release verification, regression prevention and documentation
- Design-to-code workflows: Figma analysis, shared-vs-local component classification, reusable UI mapping and implementation validation
- AI-assisted engineering: structured agent roles, reusable skills, MCP workflows, evidence-based verification and human review
I increasingly treat engineering as a system instead of a sequence of coding tasks:
Problem
↓
Investigation
↓
Architecture / boundaries
↓
Implementation
↓
Tests + performance + visual verification
↓
Documentation
↓
Release
↓
Real-world validation
↓
Reusable knowledge / tooling
AI is part of that workflow, but not the decision maker. I use agents to accelerate investigation, implementation and verification while keeping architecture, trade-offs and acceptance criteria explicit.
- Prefer evidence over assumptions.
- Treat architecture and boundaries as part of implementation, not post-work documentation.
- Automate checks that should not depend on memory.
- Separate correctness failures from conventions and heuristics.
- Ship small tools completely before expanding their scope.
- Use AI for leverage, but keep verification deterministic wherever possible.
Shipping matters more than activity counters, but consistency still matters.
- GitHub: @muhamadzolfaghari
- LinkedIn: Mohammad Zolfaghari
- npm packages: docs-healthcheck · persian-date-native
- Open-source interests: frontend architecture, design systems, developer tooling, AI-assisted engineering workflows