Open to AI / ML roles · Ready to relocate

Aliyah Alabdali

About

A model is only half the system.

I'm an Artificial Intelligence graduate from Umm Al-Qura University, with First Class Honors and a 4.0/4.0 GPA. Most of my work is in Computer Vision, NLP, and Generative AI.

My projects rarely stop at the model. I work across evaluation, APIs, interfaces, and deployment to turn models into systems people can actually use. I'm looking for an AI or Machine Learning role where I can keep working across that full lifecycle.

At a glance

Location
Saudi Arabia
Target roles
AI Engineer · ML Engineer
Availability
Available now
Projects

A selection of projects I've built.

05 case studies

Graduation Project · Computer Vision

Yaqidh: AI Child-Safety Monitoring

A real-time system that watches live CCTV, detects child falls and violence with two custom YOLOv8s models (ONNX-optimized, trained on 3,800+ collected and augmented images), and pushes zone-scoped WebSocket, SMS, and email alerts to parents, teachers, and managers, with video-backed incident records and live dashboards on a FastAPI + React stack.

Real-time child-safety monitoring that detects falls and violence with two custom YOLOv8s models and sends live, role-based alerts through a FastAPI + React platform.

0.81
mAP@50 · fall
0.66
mAP@50 · violence
2-3×
over baseline
~20%
latency ↓ (ONNX)
YOLOv8PyTorchONNXOpenCVFastAPIPostgreSQLReact
Agentic AI · Full-Stack AI

InterMind: Autonomous AI Interviewer

An autonomous AI interviewer that turns a job description into a role-specific interview, decides after every answer whether to follow up or move on, and produces an evidence-based report where every assessment traces back to something the candidate actually said. Live on Vercel with PostgreSQL on Neon, with voice through Azure AI Speech.

An autonomous AI interviewer that turns a job description into an adaptive interview, asks targeted follow-ups, and produces an evidence-based report.

0.885
Follow-up F1
239/239
Evidence Traced
3.83s
P95 Backend Latency
LangGraphOpenAIFastAPIPostgreSQLReactAzure AI Speech
Generative AI · RAG

Daleel: RAG Document Assistant

A retrieval-augmented assistant that answers questions from your own documents and names the file behind every answer. Two n8n workflows index PDFs, text, and Markdown into Supabase pgvector and run a chat agent over them; a changed file replaces its old version in a single transaction, and instructions hidden inside documents are ignored.

A RAG assistant that answers from your documents and cites its sources, built on n8n workflows, Supabase pgvector, and OpenAI embeddings.

8/9
demo checks passed
16
offline tests
n8nSupabase pgvectorOpenAIPostgreSQLRAG
Advanced Deep Learning · NLP

Nabeeh: SMS Scam Detection

An English SMS scam detector built on a Transformer encoder written from scratch in PyTorch, with custom multi-head attention, pre-LayerNorm, and masked pooling. Class-weighted training handles a heavily imbalanced corpus, and the live demo shows which words pushed each message toward scam.

A from-scratch PyTorch Transformer that flags scam SMS on a heavily imbalanced dataset, with a live demo that shows each word's influence.

98.07%
test accuracy
0.954
macro F1
Custom
PyTorch encoder
TransformersPyTorchScikit-learnGradioHugging Face
AI System Design · Privacy-first

Brain Tumor MRI Classifier

A privacy-first web app that classifies brain-MRI scans into four tumor types entirely in the browser, with no image ever leaving the device. A lightweight ~1.5 MB model runs via ONNX Runtime Web with real-time confidence scoring.

A privacy-first web app that classifies brain-MRI scans into four tumor types entirely in the browser, with no image ever leaving the device.

~79%
accuracy
~9ms
inference
~1.5MB
model size
Local
MRI processing
YOLOv8ONNX Runtime WebReactVite
Experience

Where I've worked.

Mar to Jun 2025

Artificial Intelligence Trainee (Co-op)

InPro Studio, Makkah, SA
  • Contributed to two AI-powered SaaS platforms, spanning idea classification, business-model generation, and automated document creation.
  • Applied prompt engineering to design AI workflows deployed on Azure AI Foundry.
  • Built a prompt-based classification system and co-designed a startup-readiness assessment framework scoring founders across weighted categories.
  • Authored and refined PRDs, SRSs, and BRDs, and created reusable documentation templates while supporting product design and UX improvements.
Aug 2024 to Nov 2025

Banking Trainee (Online)

Banque Saudi Fransi, Supervision Program
  • Completed a Vision 2030-aligned program in Agile, project management, and problem-solving.
Skills

The tools and technologies I use across AI development.

01

Data

PythonSQLPostgreSQL Data PreprocessingRoboflowREST APIs
02

Model

PyTorchTransformersScikit-learn OpenCVLLMs & RAGLangGraph Fine-tuning
03

Optimize & Deploy

FastAPIONNXONNX Runtime Web Model EvaluationInference OptimizationAzure Vercel

Tools & Workflow

Git / GitHubJupyterGoogle Colab Automated TestingAgilen8n

Soft Skills

LeadershipProblem-SolvingTeamwork CommunicationCritical ThinkingTime Management Self-Learning
Education & credentials

My degree and certifications.

Education & honors

BSc, Artificial Intelligence

Umm Al-Qura University · Makkah, SA · June 2026
First Class Honors GPA 4.0 / 4.0

Dean's Honor List

Academic Excellence Track · Umm Al-Qura University · Spring 2026
Certifications

STEP English Proficiency · 88/100

Education & Training Evaluation Commission (ETEC)

AWS AI Practitioner Challenge

Udacity · Issued June 2026
View credential ↗
Contact

Let's Talk.

I'm open to AI / ML roles and happy to relocate. The fastest way to reach me is email, or find me on LinkedIn and GitHub below.