Available for internships & new roles

Hi there, I'm

Omnia Ali

AI Engineer & Backend Developer

I build AI-powered systems and reliable backend platforms — from LLM pipelines and RAG evaluation to production-grade ASP.NET Core APIs and automated data workflows. Third-year Computer Science (AI Major) student at Cairo University, blending research-driven AI engineering with clean, scalable backend architecture.

Python C# / .NET LLM · RAG Azure & SQL
Omnia Ali smiling, professional portrait

01About

Cross-field engineer, one clear focus: shipping things that work.

I'm a third-year Computer Science student (AI Major) at Cairo University, currently splitting my time between two things I care about equally: building AI systems that behave the way they're supposed to, and building backend software that keeps running long after the demo is over.

That's meant evaluating LLM outputs for hallucination and bias, orchestrating multi-agent RAG pipelines, shipping ASP.NET Core APIs to Azure, automating enterprise workflows with RPA, and cleaning up messy data into dashboards people actually use. Different fields, same habit: understand the problem properly, then build the smallest reliable thing that solves it.

Location

Cairo, Egypt

Education

B.Sc. Computer Science, AI Major — Cairo University

Languages

Arabic (Native) · English (C2) · French (B2)

Currently

.NET Academy @ Luftborn · Integration Intern @ Ejada

Beyond the code: Country Ambassador for Egypt in two international competitions (IAAC & IYMC), coordinating 700+ student participants — twice recognized with the Ambassador Award for Excellent Encouragement.

02Education

Grounded in fundamentals, aimed at AI.

Omnia Ali on campus

B.Sc. Computer Science — AI Major

Cairo University · Cairo, Egypt

Expected 2027

GPA 3.7 / 4.0

Artificial Intelligence Machine Learning NLP Cloud Computing Data Structures & Algorithms Database Systems

High School Diploma

Red Sea STEM School · Hurghada, Egypt

2023

Specialized coursework & certifications

Stanford Online · DeepLearning.AI · Google · Hugging Face — see the Certificates section

03Skills

One toolbox, four disciplines.

01 · AI & LLM Engineering

LangGraphLlamaIndexRAG PipelinesPrompt EngineeringLLM EvaluationHallucination DetectionFine-tuning (LoRA)Hugging FacePyTorchTensorFlowscikit-learn

02 · Backend & Cloud

ASP.NET CoreWeb API / MVCEntity Framework CoreJWT AuthREST APIsAzure App ServiceDockerSQL ServerSQLiteRepository Pattern

03 · Data & Visualization

Pandas / NumPyETL PipelinesSQLAlchemyStreamlitMatplotlibSeabornPlotlyPower BIExcel / openpyxl

04 · Automation & Tools

RPA — Automation AnywhereHangfireGit & GitHubPostmanSwagger / OpenAPIJupyterAgile / Scrum

04Experience

Where the work actually happened.

AI Automation, Evaluation & Data Engineer

Current

Freelance · Remote — Jan 2025 – Present

  • Designed and shipped end-to-end AI-powered automation workflows for content, research, and data pipelines — cutting manual effort by over 60% across active accounts.
  • Built and iterated on multi-agent LLM systems (LangGraph, LlamaIndex), orchestrating retrieval, reasoning, and fact-checking agents.
  • Evaluated large-scale AI/ML outputs for instruction-following, factual accuracy, safety, and bias across bilingual Arabic–English datasets, using rubric-based scoring (Accuracy, Precision, Recall, F1).
  • Authored adversarial edge-case prompts and Golden Responses for alignment workflows; delivered structured evaluation reports supporting iterative model improvement.

AI Data Annotator — Bilingual Arabic–English

Mercor · Remote — Jan 2026 – Feb 2026

  • Evaluated bilingual AI-generated datasets for linguistic correctness, cultural appropriateness, and structural quality.
  • Flagged inconsistencies between source documents and model outputs within strict quality and deadline standards.

05Projects

Selected work, grouped by field.

Featured
Multi-Agent Research Assistant terminal output

Multi-Agent Research Assistant

Orchestrates specialized retrieval, reasoning, and fact-checking agents via LangGraph to answer complex multi-step questions with hallucination-aware output.

PythonLangGraphLlamaIndexRAGStreamlit
View Code
Featured
Pneumonia detection model results

MedVision — Pneumonia Detection

Benchmarked classical ML, custom CNNs, autoencoder anomaly detection, and transfer learning for chest X-ray classification; fine-tuned ViT with attention maps. Best result: ResNet50 at 95.19% accuracy, 0.9892 ROC-AUC.

PythonTensorFlowHugging FaceViT
View Code
DocuMind AI app screenshot

DocuMind-AI

PDF question-answering assistant with RAG retrieval and hallucination-aware prompting, fully deployed on Azure App Service.

PythonClaude APIRAGAzure
View Code
RAG Faithfulness Study charts

RAG Faithfulness Study

Research benchmark comparing direct prompting vs. retrieval-augmented prompting on QA tasks, evaluating faithfulness, correctness, completeness, and hallucination rate.

PythonLLM APIPandasJupyter
View Code
TestForge Cloud screenshot

TestForge-Cloud

Cloud-deployed AI service that auto-generates pytest unit tests from source code, with prompt-injection sanitization, output constraint enforcement, and AST validation.

PythonLLM APIAzure
View Code
</>

CLI AI Unit Test Generator

Command-line variant that generates deterministic pytest unit tests via LLM API, with prompt-injection sanitisation and a repair-retry validation flow.

PythonOpenAI / GeminiAST Validation
View Code
Voice Agent CLI terminal

Voice Agent CLI

End-to-end voice pipeline — mic input → Whisper STT → LLM reasoning → TTS output — with modular, swappable components and an Arabic extension.

WhisperLLM APIgTTS
View Code
Whisper WER Bench charts

Whisper WER Bench

Batch transcription pipeline across Whisper small/medium, computing per-clip and per-model Word Error Rate with CSV reports and comparison charts.

WhisperjiwerPandas
View Code
Multilingual TTS evaluation radar chart

Multilingual TTS Evaluator

Evaluated gTTS, Coqui TTS, and SpeechT5 on Arabic & English using a MOS-style rubric; visualised comparisons with bar and radar plots.

Coqui TTSSpeechT5Seaborn
View Code
Scratch vs Keras training curves

Scratch vs. Keras

Built a neural network from scratch in NumPy and benchmarked it against an equivalent Keras model, comparing training and validation loss/accuracy curves.

PythonNumPyKeras
GitHub Profile

06Certificates

Learning, documented.

Tap a certificate to view it full-size, or download the image directly — no external verification links, just the credential itself.

Supervised Machine Learning certificate

Supervised Machine Learning: Regression and Classification

Stanford Online & DeepLearning.AI · Jan 2025

Advanced Learning Algorithms certificate

Advanced Learning Algorithms

Stanford Online & DeepLearning.AI · Apr 2025

Improving Deep Neural Networks certificate

Improving Deep Neural Networks

DeepLearning.AI · Jun 2024

Google Data Analytics certificate

Google Data Analytics Professional Certificate

Google · Feb 2026

Hugging Face Agents Course certificate

Agents Course — Certificate of Excellence

Hugging Face · Apr 2025

HP LIFE Data Science and Analytics certificate

Data Science & Analytics

HP LIFE · Feb 2026

Manara Databases certificate

Databases — Learning Path

Manara · Feb 2026

IEEEXtreme 18.0 certificate

IEEEXtreme 18.0 Programming Competition

IEEE · Team Algorithm Legends · Oct 2024

IBM Critical and Creative Thinking certificate

Solving Problems with Critical & Creative Thinking

IBM SkillsBuild for Students

IBM Introduction to Emerging Technologies certificate

Introduction to Emerging Technologies

IBM SkillsBuild for Students

IBM Focus Entrepreneurial certificate

Focus: Entrepreneurial Mindset

IBM SkillsBuild for Students

University of Minnesota HR certificate

Preparing to Manage Human Resources

University of Minnesota · Sep 2021

07Contact

Let's build something intelligent.

Open to internships and full-time roles across AI engineering, backend development, and data — as well as English–Arabic technical translation and interpretation work. Have a project in mind, or think there's a fit? Reach out below or through the form.

Get In Touch

Have a project in mind or want to discuss a collaboration? Send a message and it'll open straight in your email app, ready to go.

Contact Information

Feel free to reach out through any of these channels.

Location

Cairo, Egypt

Connect with me

Download Résumé