AI & Data Engineer

Building practical
ML & data systems.

I build AI applications and the data pipelines that support them—from transaction classification and retrieval to Python services on AWS.

My work connects model evaluation, reliable data processing, and repeatable cloud deployment.

01

Professional experience

Data pipelines, analytics and applied ML.

Freelance

Data Engineer

Nov 2024 – Feb 2025

E-commerce and telecom clients

Separate client engagements

Data pipelines and analytics

  1. Cut a core 100 GB batch from 4.5 to 1.5 hours—approximately 67% less processing time on comparable workloads.
  2. Built three Spark ETL pipelines and repaired a fourth, integrating four CRM/ERP sources and processing approximately 100 GB per daily run using AWS S3, Kafka/MSK, and Airflow.
  3. Orchestrated daily Airflow runs with retries, alerts, and backfills; applied Great Expectations data validation.
  4. Designed a PostgreSQL star schema, incremental UPSERTs, and SQL window-function transformations for downstream analytics.
  5. Delivered Tableau dashboards and Python/SQL segmentation and churn analysis. Handed over DAGs, runbooks, and validation logs to client IT/Data Ops teams.
02

Selected work

Independent projects. Practical questions.

AI Multi-Agent Platform

A LangGraph supervisor coordinates customer-support agents, combining specialist responses with retrieval and SQL tools.

My contribution — Integrated Amazon Titan Embeddings V2, pgvector on Amazon RDS, and SQL correction through schema mapping and few-shot examples.

RAGAS context precision improved from 0.52 to 0.91 on a manually curated 50-pair evaluation dataset.

  • Python
  • LangGraph
  • Amazon Bedrock
  • Titan Embeddings V2
  • PostgreSQL / pgvector
  • RAGAS
How retrieval and SQL correction work

The supervisor coordinates order-management, recommendation, troubleshooting, and personalization services. Amazon Titan Embeddings V2 and pgvector on Amazon RDS support retrieval.

Schema mapping and few-shot SQL examples guide query correction. Evaluation against a manually curated 50-pair dataset measures retrieval context precision, which improved from 0.52 to 0.91.

View repository

Churn prediction system

Combines sentiment extracted from support logs with an XGBoost classifier for customer churn prediction.

My contribution — Used the DeepSeek API for zero-shot sentiment extraction across 150,000 customer profiles and fed the extracted features into XGBoost.

Independent project using a public dataset, separate from client work.

  • Python
  • DeepSeek API
  • XGBoost
03

An end-to-end perspective

01 /

AI & machine learning

Classification, forecasting, and retrieval-grounded applications, with model evaluation and experiment tracking.

  • Python
  • TensorFlow
  • PyTorch
  • XGBoost
  • TF-IDF
  • Prophet
  • RAG
  • LangGraph
  • RAGAS
  • MLflow
02 /

Data engineering

Batch and streaming ingestion, dimensional modeling, SQL transformations, orchestration, and data quality.

  • SQL
  • Spark / PySpark
  • Apache Airflow
  • Kafka / MSK
  • Apache Beam
  • PostgreSQL
  • pgvector
  • MongoDB
  • Great Expectations
  • Tableau
03 /

Backend & cloud

Python APIs, containerized workloads, and repeatable AWS delivery through infrastructure and automated checks.

  • FastAPI
  • AWS Lambda
  • ECS Fargate
  • Amazon Bedrock
  • S3 / RDS / ECR
  • Docker
  • Terraform
  • GitHub Actions
  • pytest
  • Prometheus

Certifications & learning

AI, data engineering, and cloud operations.

Amazon Web Services

AWS Certified Machine Learning Engineer – Associate

Issued September 2026

DeepLearning.AI

Data Engineering Professional Certificate

Issued December 2025

Duke University

MLOps | Machine Learning Operations Professional Certificate

Issued December 2025

LinkedIn

Kubernetes: Microservices

Issued September 2025

DeepLearning.AI

Deep Learning Professional Certificate

Issued March 2025

IBM

IBM AI Engineering Professional Certificate

Issued February 2025

Google

Google Data Analytics Professional Certificate

Issued December 2024

04 / About

Curious about the model.
Committed to the system.

I’m Mahmoud, an AI and data engineer in Cairo. I work across the pipeline, the model, and the cloud deployment.

Freelance work with e-commerce and telecom clients shaped how I deliver: validate the data, document the system, and leave a handover the client’s team can operate. My independent projects explore classification, retrieval, and AI orchestration.

Egypt · Cairo · Remote · European relocation
Languages
  • Arabic Native
  • English C1
  • German B2

05 / Let’s connect

Let’s build
something useful.

I’m looking for AI Engineer and Data Engineer roles in Cairo, remotely, or with European relocation. Let’s talk about your team’s work.

Email Mahmoud Find me on GitHub LinkedIn