Available for new projects

Full-Stack AI Engineer. LLM features, shipped end to end.

I’m Abanoub. For 7 years I’ve built production software. Today I build LLM pipelines in Python and the Node.js APIs, React frontends and cloud infrastructure that put them in front of real users.

GETapi.abanoubessam.com/v1/me
200 OK38 msapplication/json
{
  "name": "Abanoub Essam",
  "role": "Full-Stack AI Engineer",
  "based_in": "Ismailia, Egypt",
  "experience_years": 7,
  "languages": ["Arabic", "English"],
  "focus": ["LLM pipelines", "APIs", "system design"]
} 
Experience
7+ yrs

Shipping production code since March 2019.

Local time · Ismailia
--:--

Cairo time. Remote-first, overlaps with Europe and the Gulf.

How an AI request flows through my stack
Scanned PDF
Arabic & English
Page render
pypdfium2
Vision OCR
Gemini · Vertex AI
Extract & cite
Pydantic schema
Human review
approve · send
Core stack
PythonFastAPIGeminiPydanticTypeScriptNestJSPostgreSQLRedisNext.jsFlutterDockerGCP
Open source
70

Public repositories on GitHub.

AI engineering

LLM features that hold up in production

I build the Python services that turn scanned documents and open questions into structured, checked answers.

Document understanding

Vision-model OCR on scanned PDFs, then classification and field extraction into typed schemas.

{ "doc_type": "commercial_register", "issued_on": "2026-03-14", "source": "page 4, paragraph 2" }

Grounded answers

Every answer links to the page it came from, so a reviewer can check it in one click.

Evaluation

Output is compared with hand-labelled gold answers before each release.

Production plumbing

Async workers on Pub/Sub, idempotent jobs, structured logs and Docker deploys on Google Cloud. A person approves before anything is sent.

FastAPIPydantic v2google-genaiVertex AIPub/SubCloud Storageuv
Skills

What I work with every day

Grouped by layer. Every item has run in production.

AI & Python

  • Python · FastAPI · Pydantic
  • Gemini · Vertex AI
  • Vision OCR · structured output
  • Prompt design · gold-set evals

Backend

  • Node.js · TypeScript
  • NestJS · Express
  • REST · GraphQL
  • RabbitMQ · BullMQ · Pub/Sub

Data

  • PostgreSQL · MySQL
  • MongoDB · Redis
  • Elasticsearch
  • Prisma · TypeORM

Frontend

  • React · Next.js
  • Redux Toolkit · React Query
  • Tailwind CSS

Mobile

  • Flutter · Dart · Bloc
  • React Native

Infra & practice

  • Docker · Kubernetes
  • GCP · AWS · Terraform
  • System design · code review
What I build

Types of systems I’ve shipped

Described by the problem and the stack. Client details stay private.

AI · Documents

LLM document pipelines

Scanned PDFs go through page rendering, vision OCR and extraction with citations, then to a reviewer. Python workers on a queue, typed contracts, evaluation against gold answers.

PythonGeminiPub/Sub
AI · Chat

Conversational AI services

Chat APIs between a product and the model, with history, auth and rate limits.

FastAPIRedis
Backend · Payments

Booking and payment APIs

Supplier search, booking, approvals and payments, with idempotent retries.

NestJSPostgreSQL
Web · Mobile

Portals, dashboards and app backends

Role-based admin panels, customer portals, and the APIs behind Flutter apps, including moderation tools for user content.

Next.jsReact QueryFlutterMongoDB
Experience

From backend developer to AI engineer

  1. 2024 – now

    Team Lead · Full-Stack AI Engineer

    Python LLM pipelines on Vertex AI, NestJS services and Terraform-managed cloud infrastructure.

    Current
  2. 2022 – 2024

    Senior Software Engineer

    NestJS APIs, React portals, and a FastAPI chat service built with a machine-learning team.

  3. 2020 – 2022

    Senior Full-Stack Engineer

    React and Next.js frontends on my own Node.js APIs. Mentored interns.

  4. 2019 – 2020

    Backend Developer

    First production Node.js services. B.Sc. Computer Science, 2017.

Services

How I can help your team

Fixed-scope projects or a monthly retainer, in Arabic or English.

AI features

Document extraction, OCR, classification or chat added to your product, with evaluation so you know it works.

For: PDFs, support, internal tools

Backend & APIs

REST and GraphQL APIs, queues and caching that hold up under load.

For: new products, slow APIs

Full-stack products

Database to Next.js and Flutter: one engineer who owns the whole feature.

For: MVPs, dashboards

Tech leadership

Architecture reviews, code review habits, CI/CD and hiring help.

For: teams of 2–10 engineers
Open source

Code you can read

Replies within one working day

Have an AI feature to ship or an API to fix?

Send a short note about the project, timeline and budget. Arabic or English.