Mahmoud Berkoti

(

(P.A.I.D) - Personal AI Doctor

• Developed a web-based application for medical diagnosis using advanced AI techniques, including LLMs and RAG
• Implemented a user-friendly interface with HTML, CSS, and JavaScript, and built the backend with Python and Flask for secure API communication
• Integrated LLAMA 3 70B to convert user input into keywords and Mixtral 8x7B to process information from vector databases for accurate responses
• Optimized the system for real-time, relevant AI-generated medical advice, ensuring easy customization and deployment on any infrastructure
Skills: Python (Programming Language)
• Large Language Models (LLM)
• JSON
• Mixtral 8x7B
• JavaScript
• Machine Learning
• Application Programming Interface

I

IntelliVulnScan

• Developed a vulnerability management platform for 12 test environments using React, Node.js, Express,
Prisma, and SQLite, enabling real-time detection and reducing manual oversight by 60%.
• Integrated a scanning engine with Node.js to identify critical risks in under 2 minutes, maintaining an 85%
detection rate across mock datasets.
• Built a React-based dashboard featuring 5 interactive charts, providing immediate insights into high-risk assets and decreasing incident response times by 30%.
• Implemented secure JWT authentication while leveraging Prisma and SQLite to keep average response times below 1 second during local load tests.
Skills: Node.js
• Threat & Vulnerability Management
• JWT
•

A

AI Training Bot

• Developed an AI trading bot using Python and Alpaca API integrated with Lumibot and
YahooDataBacktesting
• Implemented real-time sentiment analysis using FinBERT and PyTorch, enhancing financial decision-making with 99% accuracy
• Utilized REST API calls to Alpaca for data retrieval and trade execution, ensuring swift market responsiveness
• Achieved a remarkable 68.56% annual return on $SPY in back testing, demonstrating the strategy’s effectiveness
Skills: FinBERT
• REST APIs
• PyTorch
• Alpaca API
• Lumibot

M

Motion Detection Security System (MDSS)

• Developed a Python-based security camera system using OpenCV for advanced image processing and Tkinter for a user-friendly frontend.
• Optimized motion detection algorithm to achieve an 85% accuracy rate in highlighting and tracking motion.
• Integrated local video storage with an automatic recording stop mechanism, reducing storage usage by over 30% during non-motion periods.
• Delivered a fully integrated security solution, combining real-time video processing, intuitive interface design, and efficient storage management.
Skills: Tkinter
• OpenCV
• Object-Oriented Programming (OOP)

D

DataWhisk

• Automated Review Collection – DataWhisk streamlines the process of gathering Google reviews using Playwright, eliminating the need for manual data collection and ensuring a continuous flow of customer feedback.
• AI-Powered Sentiment Analysis – By leveraging advanced transformer models, the platform analyzes customer sentiments, identifying trends and key emotions in reviews to help businesses make data-driven decisions.
• Insightful Data Categorization – Reviews are automatically sorted into key categories like food, service, and pricing, providing businesses with a clear understanding of what customers appreciate and where improvements are needed.
• Visualized Insights for Decision-Making – Using tools like Matplotlib and

E

E-commerce Website

• Robust Structuring and Accessibility: Our website leverages HTML for structured content, ensuring accessibility and SEO friendliness, while CSS provides a visually appealing and consistent design across various devices.
• Enhanced Interactivity and Compatibility: JavaScript enhances user interaction with dynamic elements like sliders and pop-ups, ensuring smooth functionality across all major browsers including Chrome, Firefox, Safari, and Edge.
• Optimized Performance: CSS optimizations improve load times and site responsiveness, ensuring a seamless user experience regardless of device type.
• User Experience and SEO Focus: The site design prioritizes user experience with intuitive navigation and aesthetic appeal, coupled

A

Aim Training Simulator

• Developed an aim training simulator using Python and Pygame library with Pygame GUI for the user interface
• Implemented dynamic target generation with varying sizes and speeds, providing a challenging shooting experience
• Incorporated power-up features including extra lives, slow motion, and larger targets to enhance gameplay strategy and diversity
Skills: Python (Programming Language)
• Object-Oriented Programming (OOP)
• Graphical User Interface (GUI)
• Pygame
• Pygame GUI

About

Detail

Timeline

Résumé

Mahmoud Berkoti

About

Detail

Behavioral traits