Sep 2022 - Jul 2026
B.Sc. Computer Science & Artificial Intelligence
Ain Shams University & University of East London dual degree in Cairo, Egypt. GPA: 3.3 / 4.0.
Graduated: July 2026
AI Engineer | Computer Vision | NLP | Healthcare AI
Hi, I'm Omar Salama
AI Engineer building production computer vision and NLP systems - from a 96.69% AUROC real-time seizure detection pipeline to deployed healthcare AI. CS & AI dual-degree graduate from Ain Shams University and the University of East London.
About Me
Turning applied AI research into production-ready systems for high-impact applications.
AI Engineer focused on Computer Vision, NLP, and Real-Time Inference Systems
I build AI systems that bridge cutting-edge research with real-world deployment, specializing in computer vision, natural language processing, and performance-aware inference. I'm a dual-degree graduate in Computer Science and Artificial Intelligence from Ain Shams University and the University of East London (Class of 2026, GPA: 3.3/4.0).
My flagship work is Shifaa, an AI-powered, privacy-first clinical decision support and patient monitoring platform. I engineered its real-time seizure detection pipeline using Vision Transformers, Graph Neural Networks, OpenPose, ONNX Runtime, C++, and Python, achieving 96.69% AUROC, 90.18% F1-score, and 37+ FPS for low-latency pre-clinical seizure detection.
Beyond healthcare AI, I've contributed to NLP research at MAHE, autonomous path planning with the ASU Racing Team, earned the AWS Certified Machine Learning – Specialty certification, and mentored aspiring AI engineers through ApplAi. I'm currently applying that same engineering mindset to AI automation at Cultiv Bureau, building n8n, RAG, CRM, and API-driven workflows.
Computer Vision
NLP
Healthcare AI
Research
Leadership
Experience & Education
Sep 2022 - Jul 2026
Jul 2026 - Present
AI & Automation Intern | Cultiv Bureau
Build AI-powered business automation workflows using n8n, REST APIs, webhooks, WhatsApp Cloud API, and Bitrix24 CRM integrations.
Focused on RAG pipelines, prompt optimization, token reduction, Docker-based local development, building multi-agent systems, and production-style debugging.
Jul 2025 - Jan 2026
Data Science Trainee | DEPI
Completed intensive cohort-based training in Python, machine learning, deep learning, data visualization, and AI deployment workflows.
Applied supervised and unsupervised learning across end-to-end data science projects.
Jul 2025 - Sep 2025
Data Scientist Intern | MAHE
Built and deployed an interactive news summarization chatbot using Google Gemini API with conversational, context-aware question generation.
Co-authored an NLP text summarization research paper under Prof. Ramakrishna.
Oct 2022 - Aug 2025
NLP Specialist| ApplAi
Developed and delivered a hands-on Deep Learning and NLP curriculum focused on TensorFlow, Transformer architectures, and modern AI workflows for 50+ participants.
Oct 2024 - Aug 2025
IT Specialist | Dispute
Managed DNS infrastructure and executed large-scale email automation campaigns, improving system reliability and streamlining organization-wide communication pipelines.
Oct 2024 - Jul 2025
Path Planning Specialist | ASU Racing Team
Developed a path-planning system for an autonomous vehicle, implementing optimization algorithms from scratch for the Shell Eco-marathon team.
Jul 2024 - Oct 2024
Full-Stack Developer | Notopia
Built a scalable e-commerce app using Flutter, Next.js, and SQL. Optimized backend queries, reducing processing time by 30%.
Featured Projects
Flagship AI systems presented as compact case studies for fast technical review.
Graduation Project | Healthcare AI
Shifaa - AI-Based Patient Monitoring Platform
Production-grade, privacy-first hospital monitoring platform unifying a grounded medical AI assistant with real-time detection for cardiac arrhythmia, patient falls, and epileptic seizures.
96.69% AUROC
90.18% F1
37+ FPS
ProblemHospitals need fast, on-premises patient monitoring without leaking sensitive data.
SolutionA role-aware interface with three detection pipelines and a grounded medical assistant.
ArchitectureSeries-gated ViViT + GNN seizure pipeline with OpenPose, ONNX Runtime, C++, and Python.
Results96.69% AUROC, 90.18% F1-score, 37+ FPS, and pre-clinical seizure alerts.
ImpactDesigned for privacy-first, fully on-premises hospital deployment.
Future WorkClinical validation, alert workflow hardening, and broader hospital integrations.
1st Place | CV Course Competition
Fruit & Food Segmentation and Calorie Estimation
AI-powered food analysis platform integrating segmentation, classification, few-shot recognition, and calorie estimation with an EfficientNet-B4 + DeepLabV3+ model.
97.69% Val IoU
1st Place
ProblemFood analysis needs object-level recognition before calories can be estimated reliably.
SolutionSegment food regions, classify items, support few-shot recognition, then estimate calories.
Results97.69% Validation IoU and 1st Place in the Ain Shams CV Course Competition.
LessonsStrong segmentation quality is the foundation for trustworthy downstream nutrition estimates.
Medical Imaging
Liver Tumor AI - Real-Time Detection & Segmentation
PyTorch + U-Net liver tumor segmentation pipeline for unseen patients using strict patient-level validation.
95% Accuracy
91.7% Dice
ProblemMedical segmentation can look strong while leaking patient data across splits.
SolutionPatient-independent validation with a U-Net segmentation pipeline.
Results95% accuracy and 91.7% Dice score.
ImpactHealthcare AI proof point with a deployed demo and reproducible code.
NLP Research
Article Summarizer Chatbot
Conversational summarization and question-generation chatbot using Gemini 2.5 Flash API.
ProblemReaders need fast, contextual summaries and follow-up questions from long articles.
SolutionA Streamlit chatbot that summarizes, extracts topics, and generates context-aware questions.
ResearchBuilt during the MAHE internship and connected to an NLP summarization paper.
Future WorkAdd richer evaluation dashboards and source-grounded answer tracing.
NASA Space Apps
ExoNet - AI-Powered Exoplanet Classifier
Residual MLP classifier for NASA K2 exoplanet candidates with SHAP explainability.
95.28% Accuracy
ProblemCandidate signals need interpretable classification to support review.
SolutionResidual MLP classifier paired with SHAP feature attribution.
Results95.28% accuracy in the NASA Space Apps Challenge context.
ImpactShows ML breadth beyond healthcare, with explainability in a scientific workflow.
Technical Skills
CV-aligned skill groups for technical recruiters, ATS scans, and fast engineering review.
Programming Languages
Python
C++
C#
JavaScript
Dart
SQL
AI & Machine Learning
PyTorch
TensorFlow
Scikit-learn
Hugging Face Transformers
ONNX Runtime
Graph Neural Networks
Vision Transformers
RAG
Prompt Engineering
Computer Vision & NLP
OpenCV
Image Segmentation
Object Detection
Pose Estimation
Optical Flow
3D Vision
Transformers
Gemini API
Data & Cloud
Pandas
NumPy
AWS SageMaker
AWS Lambda
Data Preprocessing
Streamlit
n8n
REST APIs
Webhooks
Bitrix24 CRM
WhatsApp Cloud API
Multi-agent Systems
Software Development
Flutter
Next.js
Git
Linux
Docker
OOP
Data Structures & Algorithms
Leadership & Research
Technical Leadership
Cross-functional Collaboration
Public Speaking
Problem Solving
Curriculum Design
Research Writing
Achievements & Certifications
Amazon Web Services | May 2024
AWS Certified Machine Learning - Specialty
Validated machine learning model building, training, tuning, and deployment on AWS, including SageMaker-centered workflows.
Graduation Project Expo
Shifaa Participated in ASU GP Expo
Presented the AI patient monitoring platform to faculty, industry professionals, and recruiters at the FCIS Expo.
NASA | Sep 2025
NASA Space Apps Challenge Participant
Built ExoNet, an explainable exoplanet classifier reaching 95.28% accuracy on K2 candidates.
Huawei | Nov 2023
HCIA-AI V3.5
Certification covering AI foundations, neural networks, deep learning, and cloud AI concepts.
NTI | Nov 2023
AI Professional Training - 120 Hours
Government-backed intensive program in ML, DL, ANN, data analysis, and professional skills.
ApplAi | Jul 2023
Computer Vision Professional Training
Hands-on training in object detection, image classification, and OpenCV pipelines.
ICPC | Jun 2023
ICPC ECPC Qualification
Competitive-programming qualification demonstrating algorithmic problem-solving fundamentals.
ArabCAB 2026
FCIS Team Finalist
Represented the Ain Shams University FCIS team as a finalist at ArabCAB 2026.
EF SET | Oct 2025
English C1 Advanced
Advanced English proficiency for technical writing, presentations, and cross-functional collaboration.
Leadership & Community
Leadership roles that support the AI engineering narrative through mentoring, partnerships, and community delivery.
Oct 2022 - Aug 2025
ApplAi | Operational Lead
Led 12+ strategic meetings, secured 3 partnership and venue agreements, and delivered 20+ soft skills workshops to develop members' professional and leadership capabilities.
Mar 2023 - Present
Resala Charity | Head of Department
Directed 50+ volunteers and organized 10+ charity events, including the El-Adha Clothing Campaign 2023, across 30+ social initiatives.
Sep 2025 - Apr 2026
IAESTE | Exchange Administrator
Managed the 2025-2026 international internship exchange cycle, coordinating internship placements across multiple countries.
Jan 2025 - Nov 2025
AIESEC | Finance & B2B Specialist
Conducted market research, closed 2 strategic partnership deals, and optimized financial workflows across university entities.
Get In Touch
omarsalama117@gmail.com
Phone
+20 102 498 0973
omar-mohamed-salama
GitHub
omarsalama4
Location
Cairo, Egypt