Full SaaS products, idea to revenue.
Next.js, Supabase and Stripe: auth, payments, database and the whole product loop, shipped solo from first commit to live and charging.
- Next.js
- Supabase
- Stripe
- Postgres
Available for a new build · Summer 2026
I build AI products that ship to real users.
Full-stack engineer. I ship SaaS products with applied AI on Next.js, Supabase and AWS, from the first Figma frame to production in the same week.
- Next.js
- TypeScript
- Supabase
- Stripe
- OpenAI & Anthropic
- RAG & Embeddings
- AWS Lambda
- Postgres
- Terraform
- Docker
- React Native
- GitHub Actions
What I ship
LLM features wired into real workflows.
RAG pipelines, structured outputs, evals, built into the product, not bolted on the side.
- Anthropic
- OpenAI
- Embeddings
AWS that doesn't fight you.
Lambda, DynamoDB, S3, IaC, and CI/CD that takes a commit to production without ceremony.
- Terraform
- Actions
- Observability
Figma frames into shipped UI in days, not sprints.
One engineer who can read a design file, push back where it matters, and turn it into the real, accessible, fast thing.
- Design systems
- Motion
- a11y
React Native apps that feel native on the device.
From the first onboarding screen to App Store and Play Store release, TypeScript, Expo, native modules where it counts, OTA updates so you can ship on a Tuesday.
- iOS
- Android
- Expo EAS
- Push & Deep links
How a build goes
-
01
Frame the problem
One call, then a written brief. We agree on the smallest thing worth shipping and what success looks like before any code.
-
02
Design the surface
Figma frames, copy passes, and the unglamorous edge cases. Done so the build is mostly typing, not deciding.
-
03
Build in the open
Daily commits, weekly demos, a staging URL you can poke at. No mystery, no big reveal.
-
04
Ship & hand off
Production deploy, monitoring, a runbook, and a build your team can keep extending after I'm gone.
The stack I reach for
Mobile
- React Native
- Expo / EAS
- Swift / Kotlin (when needed)
- Reanimated · Skia
Web
- TypeScript
- Next.js / Vite
- Tailwind · SCSS
- Node / tRPC
AI
- Anthropic Claude
- OpenAI
- RAG · pgvector
- Evals · structured output
Cloud
- AWS · Lambda · DynamoDB
- Postgres · S3 · CloudFront
- Terraform · GitHub Actions
- Sentry · OpenTelemetry
About
I'm Abdel, a full-stack engineer working across product, AI, and infrastructure. I spent nine years on a production e-commerce platform serving real traffic, which is where I learned to ship things that hold up. These days I build and ship SaaS products of my own, folding LLMs and RAG in where they change how people use the product. I build React Native apps too.
I work best with teams who care about craft and want one engineer who can take a thing from a Figma frame to a Docker container in the same week, no hand-offs, no theatre.
Writing
Applied AI
Stop parsing LLM text. Use tool calls.
Asking a model for JSON and parsing the text breaks in production. Define a typed tool, force the model to call it, and get validated structured data back.
Read →
Next.js
Your Next.js Server Action is a public endpoint
A server action looks like a function call, but it compiles to a public POST endpoint anyone can hit with any arguments. Authenticate and validate inside it.
Read →
Scale
What a high-traffic checkout taught me about reliability
Years on a production e-commerce checkout came down to a few hard rules. Make every operation idempotent, never trust the network, and degrade instead of failing.
Read →Get in touch
Project work, AI questions, or something you want to build. I read every email and reply within a couple of days.
abdel@codefolio.dev