Full Stack AI Developer | Specializing in Deep Learning and Software Engineering
I can help you make your system intelligent from design ,math to deploy 🐰🤖.My Story ->
Designed a smart traffic signal system using Reinforcement Learning, trained on simulated Egyptian road networks. Applied experimental design and ANOVA to evaluate models. Deployed the top-performing model in a prototype with real-time Computer Vision for speed estimation. Integrated IoT for signal control and developed a mobile app for emergency alerts. Project released under MIT license. Following Image is for final prototype
Developed an AI-powered assistant for programmers using a microservices architecture (FastAPI, Node.js), processing 1K+ daily requests via a Redis-based queue system with 3 consumers. Implemented task-based agent logic (5 states) with RAG and LLM integration, improving response accuracy by ~30%. Enhanced performance with caching, connection pooling, and CDN, reducing latency by ~40%. Integrated monitoring (Sentry, UptimeRobot) and secure authentication (Firebase JWT) to ensure reliability and scalability.
A cross platform application for providing first aids, using Health data and LLMs.
Analyzed key economic indicators, including GDP growth in Egypt from 1995 to 2018, and compared them with Africa, Saudi Arabia, and global averages. The analysis found that Egypt’s average GDP growth was approximately 5.3%, using Power BI and Excel for data analysis and visualization.
Performed real-time social media analysis, diagnosing frequent user complaints using Kafka, Reddit, Python, and GPT, achieving 70% accuracy.
An Android application for removing or replacing image backgrounds and applying ToF camera effects using computer vision. Built with YOLO, Android, FastAPI, and HttpClient. Project released under MIT license.
Lead developer of an Android mini file explorer application; designed and developed the app, with compression features mostly handled by the team. Project released under MIT license.
Created a Python pip package for SVM classification, implementing hinge loss and subgradient descent. Project released under MIT license.
Developed an AI-driven Tic-Tac-Toe game using Java and Python, implementing the alphabeta pruning algorithm. Project released under MIT license.
Built a Windows CLI app for symbolic differentiation, focusing on design, recursion, and data structures in Dart.
Studied core concepts of virtualization, fault tolerance, and cloud service models including IaaS, PaaS, and SaaS. Gained strong theoretical understanding of how virtual machines, hypervisors, and distributed systems ensure availability and scalability. Built academic-level knowledge of cloud architecture design principles and service deployment models without direct production implementation..
Led the AI team in designing and developing intelligent systems focused on cost optimization and performance efficiency. Architected scalable AI solutions that balanced computational cost with high model accuracy and fast inference time. Collaborated with cross-functional teams to align AI strategies with business goals and measurable performance improvements..
Recognized and personally recommended by senior management for outstanding technical skills and strong commitment to excellence. Demonstrated exceptional problem-solving ability and innovation in AI and software development. Acknowledged for delivering high-impact results and contributing significant value to the project’s success.
I am Taha, a Computer and Data Science graduate with an Excellent degree. From my first year at university, I focused on building a strong AI career path by following structured roadmaps, earning certifications, and developing real-world projects. Throughout my academic journey, I studied a wide range of subjects — from advanced mathematics and statistics to software engineering and system design — and worked on deeply understanding them. This enabled me to build strong foundations and develop the ability to design and implement end-to-end AI systems, from data processing to deployment.
