Professional Journey

Managed cross-functional product pod (5 engineers, 2 designers) to deliver LLM-driven features co-designed with clinicians

Engaged 25+ stakeholders to translate pain points into LLM product specs, accelerating roadmap by 30%

Integrated generative-editing LLM that reduced note-writing time by 65% with 92% clinician satisfaction

Built HIPAA-compliant PHI anonymizer processing 2M+ tokens/day with 99.3% entity recall

Developed appointment-volume predictor reducing overstaffing costs by 12% (MAPE 6.4%)

Led fine-tuning and evaluation of domain-specific LLMs improving factual accuracy by 8 percentage points

Automated workflows in radiation oncology using tree-structured LLMs

Engineered anonymization pipeline for VARIAN and MOSAIQ databases

Built clinical NLP pipeline predicting diagnosis types and service levels from unstructured consult notes

Implemented advanced dataset synthesizer using transformers boosting security model accuracy by 5%

Achieved state-of-the-art results with 81.1% accuracy, beating the DoppelGANger model

Focused on security and machine learning research, specifically using ML for enhanced security

Taught CS 2505 (Computer Organization in C and Assembly) to 432 students with excellent evaluations

Instructed CS 4114 (Formal Languages and Automata Theory) for 225 students

Maintained high teaching standards with strong student feedback and engagement

Built ASR captioning system using Wav2Vec2.0, reducing turnaround time by 90%

Deployed production models with Flask & Docker serving 10K+ weekly requests

Mentored 50 students in deep learning bootcamp; 100% completed capstones

Developed AI-powered Automated Assistant Tool to enhance financial inclusion and accelerate digitization

Achieved 6th place among national finalists in competitive innovation challenge hosted by Emirates NBD

Enabled contextual banking services through advanced AI implementation

Developed OCR system for recognizing text in old Arabic books using generative models

Built system to capture, classify, and simulate Arabic complicated fonts

Learned to build Convolutional Neural Networks (CNNs) for object detection

Launched student activity at Cairo University teaching programming fundamentals with C++ focus

Trained 600+ students in problem-solving methodologies and programming best practices

Built comprehensive curriculum covering basic programming concepts to advanced problem-solving