Let's Talk ↗
Available for work

Hi, I'm Ahmed Talaat. AI Engineer

AI Engineer | Full Stack AI Engineer I turn your raw, messy data into production-grade AI systems that predict, automate, and scale. Full-stack, explainable, and cloud-ready.

Scroll
Python
PyTorch
TensorFlow
Keras
Scikit-Learn
HuggingFace
Pandas
NumPy
Apache Spark
XGBoost
LightGBM
NLTK
OpenCV
Docker
Kubernetes
AWS
MS Azure
MLflow
Airflow
DVC
PostgreSQL
MongoDB
FastAPI
Streamlit
Power BI
Plotly
SQL
Git
GitHub Actions
SHAP
Matplotlib
SQLite
scrapling
pdfplumber
Tesseract OCR
Python
PyTorch
TensorFlow
Keras
Scikit-Learn
HuggingFace
Pandas
NumPy
Apache Spark
XGBoost
LightGBM
NLTK
OpenCV
Docker
Kubernetes
AWS
MS Azure
MLflow
Airflow
DVC
PostgreSQL
MongoDB
FastAPI
Streamlit
Power BI
Plotly
SQL
Git
GitHub Actions
SHAP
Matplotlib
SQLite
scrapling
pdfplumber
Tesseract OCR
(01) — About Me

You have data. Maybe a lot of it. But raw data doesn't make decisions — intelligence does.

I build the systems that bridge that gap. Whether you need to predict what happens next, understand why customers leave, automate your most expensive workflows, or deploy an AI product your users will actually love — I design and build it end-to-end.

No handoffs. No jargon. Just working intelligence, shipped fast.

0 AI Projects shipped
0 Tools mastered
NASA Space Apps Challenge
Data never lies
(02) — What I Offer

End-to-end AI
capabilities.

Strategy, data, models and deployment — all under one roof.

01

Machine Learning & Predictive Modelling

XGBoost · Deep Learning · Time-Series · SHAP
02

Data Analysis & Business Intelligence

EDA · Dashboards · Power BI · Tableau
03

Natural Language Processing

Sentiment · Classification · HuggingFace · LLMs
04

MLOps & Pipeline Automation

MLflow · Airflow · DVC · Docker · Kubernetes
05

Full-Stack AI Application Development

FastAPI · Streamlit · Node.js · AWS · Azure
06

Computer Vision & Image Intelligence

OpenCV · CNNs · PyTorch · Object Detection
(03) — Expertise

How I solve your
AI challenges.

Architectural depth across the full ML engineering stack.

🧠

AI & ML Engineering

  • Predictive systems with XGBoost & K-Means for real-world business applications
  • Deep learning models in TensorFlow & PyTorch for spatial and temporal data
  • Explainable AI (SHAP) frameworks for model transparency & stakeholder trust
  • Hugging Face ecosystem for state-of-the-art NLP & generative AI
  • Computer vision solutions with OpenCV at production scale
📊

Data Science & Analysis

  • Raw datasets into actionable insights — Pandas, NumPy, Spark, Hadoop
  • Interactive dashboards with Streamlit, Power BI, and Tableau
  • Comprehensive EDA and statistical modeling for strategic decisions
  • Data warehouse solutions: PostgreSQL, BigQuery & Snowflake
  • Advanced regression, clustering & time-series forecasting
⚙️

DevOps & MLOps

  • CI/CD pipelines with GitHub Actions & Jenkins for rapid deployment
  • AI microservices containerised with Docker & Kubernetes
  • Model lifecycle management & version control — MLflow & DVC
  • Multi-stage ETL workflows orchestrated with Apache Airflow
  • High-performance ML architectures on AWS & Azure cloud
(04) — Selected Work

Projects that
move the needle.

All projects ↗
01
Featured · Machine Learning

E-Commerce Churn Predictor Pro

Enterprise-grade ML solution predicting high-value customer churn before it happens — utilising XGBoost, K-Means clustering, and SHAP explainability within an interactive Streamlit dashboard that gives business teams real-time, actionable risk signals.

PythonXGBoostStreamlitPandasSHAPK-Means
02
Space & AI · NASA Challenge

CTRL-Sky Weather Prediction System

Advanced weather prediction system built for the NASA Space Apps Challenge — using ML and massive geospatial datasets for high-precision Earth observation forecasts, deployed on AWS with automated data ingestion from NASA satellite APIs.

PythonTensorFlowGeospatialAWSOpenCVNumPy
03
Healthcare AI · ML Dashboard

SynapseMed — Clinical Diagnostic AI

Advanced ML diagnostic dashboard with segregated Patient & Doctor portals — predicts 48 diseases with 99.5% accuracy from symptoms, ADA-compliant diabetes agent with PDF/image OCR, non-invasive risk scoring, bilingual EN/AR UI, and a live clinic management suite.

PythonStreamlitScikit-learnPandaspdfplumberTesseract OCR
04
Web Scraping · Data Engineering

Amazon Data Scraper & Analytics

Resilient, testable web scraping pipeline for Amazon search results — using browser-fingerprint impersonation (scrapling + curl_cffi) to bypass bot detection, storing enriched product data in SQLite with CSV export, and generating visual analysis reports (price histograms, ratings scatter plots) via Pandas & Matplotlib.

PythonscraplingSQLitePandasMatplotlibStreamlit
(05) — My Process

From raw data
to live impact.

A clear, structured process so you always know what's coming next.

STEP 01

Discover

We dig into your data, business goals, and what "success" actually means — before writing a single line of code.

STEP 02

Analyse

Deep exploratory data analysis to uncover patterns, clean noise, and identify the most powerful predictive signals.

STEP 03

Build & Train

Select, tune, and validate the right model architecture. Every decision is explainable and benchmarked rigorously.

STEP 04

Deploy & Scale

Package into production-ready APIs, containerise with Docker, and deploy to cloud — with monitoring baked in from day one.

(06) — FAQ

Questions you
probably have.

I work across the full ML spectrum — predictive modelling (classification, regression, time-series forecasting), NLP (sentiment, classification, LLM integrations), computer vision, and end-to-end MLOps pipelines. If you have a data problem with a business outcome attached, I can build the intelligence layer around it. Industries include e-commerce, finance, healthcare data analysis, environmental science, and SaaS products.
A focused ML model with clean data (e.g. a churn classifier, sentiment analyser, or recommendation engine) typically takes 2–4 weeks end-to-end — from EDA to a deployed API. Larger pipelines with messy data ingestion, MLOps setup, and cloud deployment can take 6–10 weeks. I always share a clear milestone plan upfront so you know exactly what's happening and when.
Absolutely — that's my core strength. I handle the entire lifecycle: data cleaning & EDA → feature engineering → model training & evaluation → explainability (SHAP) → containerisation (Docker) → cloud deployment (AWS / Azure) → monitoring. No handoff gaps, no miscommunication between teams. You get one engineer who owns the full stack.
Yes — always. I collaborate with businesses and founders globally using async communication, shared notebooks, live dashboards, and scheduled video check-ins. Time zones are never a blocker. I'm currently based in Egypt and have worked on projects involving NASA datasets (US), geospatial data, and international e-commerce platforms.
Just three things: (1) a description of your problem and what success looks like, (2) access to your existing data (even if it's messy — that's normal), and (3) a quick discovery call. I'll handle the rest with a structured project plan, data audit, and timeline delivered within 48 hours of our first conversation.
I build with production in mind from day one. Every model goes through rigorous cross-validation and hold-out testing. I use MLflow for experiment tracking, DVC for data versioning, Docker for reproducibility, and set up automated retraining pipelines with Airflow. I also instrument models with drift detection so you're alerted the moment performance degrades in the real world.
(07) — Contact

Let's build
something intelligent.

Most businesses sit on a goldmine of data but don't have the shovel. I build custom AI tools that turn your raw numbers into actionable growth. Let's turn your "what ifs" into "what's next."

Location
Egypt · Available worldwide
Response time
Within 24 hours
Status
✓ Available for new projects
or