Senior full-stack engineer.
Next.js · commerce · agentic AI.
I write the code that ships to production: App Router storefronts, headless commerce on Shopify and commercetools, and TypeScript services with real observability, shipped through agentic workflows with Claude Code, Codex, MCP servers, custom skills and subagents.
Four capabilities. Picked deliberately.
Three areas where I'm genuinely senior, and one way of working that runs through all of them. Everything else is a tool I'll reach for when it's the right one.
Next.js & React, end-to-end
App Router, RSC, edge routing, ISR, and Core Web Vitals tuned against real-user data.
Headless & composable commerce
Shopify, Adyen, Contentful, Amplience.
Backend & API engineering
Node.js, FastAPI, GraphQL, REST, event-driven services.
Agentic engineering
Claude Code, Codex, MCP servers, custom skills and subagents wired into build, review and ship.
See the full stack ↓
Production work, ranked by what mattered.
Brands you've heard of. Measurable outcomes. Architecture that scales.
commercetools
Composable, headless commerce APIs powering enterprise e-commerce storefronts.
Elevate AB
Shopify A/B testing and experimentation platform with storefront-first architecture and monorepo tooling.
M&M'S
Global M&M'S direct-to-consumer storefront on a headless commercetools stack. Next.js frontend, Amplience content, Adyen payments, production Core Web Vitals.
Mars, Incorporated
Mars, Incorporated corporate brand site. Next.js + Contentful headless architecture at enterprise scale.
Ocucon
AI-powered computer vision and video analytics platform. React dashboards over a Python/OpenCV pipeline.
Gradion
A global digital consultancy driving innovation and transformation across industries.
Agents I've built, and how I ship with them.
Skills, subagents and assistants running in real teams. Evidence over buzzwords.
A release-hotfix agent for an enterprise platform team
Cutting a hotfix on a multi-component platform meant a checklist of manual steps across two repositories. I packaged the whole sequence as a Claude Code skill backed by a dedicated subagent: point it at a component, an environment and the fix PRs, and it resolves what's actually deployed, builds the hotfix branch, tags it, and raises the environment bump PR for review.
Natural-language time tracking on Kimai
An assistant for Gradion engineers that turns sentences like "from 9 to 11:30 I worked on the login bug" into correct timesheet entries in Kimai. Built entirely as a CLAUDE.md-driven Claude Code workflow: onboarding, project lookup, activity mapping and confirmations, with no custom app to deploy or maintain.
How I ship: Claude Code, Codex and MCP in one loop
The setup behind my day-to-day delivery. Claude Code is the primary agent, Codex is the second opinion and rescue path when a problem needs a fresh diagnosis, MCP servers connect both to the systems I actually work in, and parallel subagents handle broad searches and codebase-wide changes. I review every diff that ships.
Questions people ask before booking.
Who are you?
I'm Ibrahim Aly, a senior full-stack software engineer. I've shipped Next.js App Router storefronts, Shopify A/B testing platforms, and commercetools work for brands including M&M'S, Mars, DEPOT, and Ocucon. I also build Claude Code skills, subagents and MCP-connected workflows that teams use to ship.
What do you specialize in?
Four things. (1) Next.js full-stack: App Router, React Server Components, TypeScript, Vercel. (2) Headless commerce: commercetools, Shopify, composable architectures, Core Web Vitals on real storefronts. (3) Backend & API engineering: Node.js, FastAPI, GraphQL/REST services, AWS, observability. (4) Agentic engineering: Claude Code, Codex, MCP servers, custom skills and subagents built into the delivery workflow.
How can you help my business?
If you need to migrate to Next.js App Router, stand up a headless commercetools or Shopify storefront, build production backend services, or fix Core Web Vitals on a revenue-critical site, I own the delivery end-to-end and ship.
What's your development approach?
Clean TypeScript, tight feedback loops, real tests (Playwright, Jest), monorepo-friendly (Turborepo, pnpm workspaces), and production-grade CI/CD from day one. I build for the team that maintains it, not just the demo.
Do you let AI write code you don't understand?
No. Agents draft; I own the diff. I use Claude Code and Codex daily, with MCP servers, custom skills and subagents, to move faster on the parts that are well understood. On commerce-critical paths like checkout, pricing and payments I read every line, and agent output that passes its tests but misreads a business rule gets rewritten, not merged.
How do we start working together?
Send a short brief, scope, timeline, what's blocking you. I'll respond within 24h with a plan, a price, and the first call booked.