KHALED ABDELKARIM

Motivated programmer proficient in Python, C, C++, and NodeJS, with hands-on experience in frameworks and technologies such as ROS2, LangChain, LlamaIndex, LangGraph, and Socket.IO. Adept at developing robust systems for natural language processing, autonomous navigation, voice-based interactions, and multi-agent architectures. Excels in delivering scalable and efficient solutions tailored to dynamic business needs. An expert troubleshooter and critical thinker with advanced problem-solving abilities, specializing in autonomous systems, AI-driven chatbots, and real-time communication platforms.

• Silver Medal in Malaysia Startup Challenge 2025, Intelligent Drone Safety and Recovery System with Advanced Diagnostic Analytics and Emergency Response Mechanisms.

Autonomous Drone Exploration System Dec 2024 — July 2025

• Objective: Design and implement a complete, end-to-end software stack for a drone to perform fully autonomous exploration and mapping in unknown, GPS-denied environments.
• Details: Designed a hierarchical perception and planning architecture using ROS1. Integrated an OAK-D camera with RTAB-Map for real-time Visual-Inertial SLAM and localization. Developed a dynamic 3D world model by integrating packages such as map_manager and onboard_detector to track moving objects and maintain map purity. Implemented a two-layer motion planner, utilizing a Dynamic Exploration Planner (DEP) for high-level frontier detection and a Polynomial Trajectory Optimizer to generate smooth, dynamically feasible flight paths. Created a web-based mission control dashboard using rosbridge for real-time visualization and supervisory control.

Autonomous Customer Service Robot Jun 2024 — Nov 2024

• Objective: Build an autonomous robot for customer service with advanced navigation and interaction capabilities.
• Details: Utilized ROS2, SLAM, and Nav2 for dynamic mapping, localization, and obstacle avoidance. Integrated voice and visual recognition systems to engage customers, provide product recommendations, and handle appointments. Developed autonomous docking features to support future upgrades, such as charging station integration.

Chatbot Development Jun 2024 — Nov 2024 

• Objective: Develop a multi-tenant chatbot with personalized interactions for diverse clients.
• Details: Used LangChain and LlamaIndex for chatbot orchestration and retrieval-augmented generation (RAG) for dynamic, context-aware responses. Created separate models for unique use cases, such as pet travel regulations and product inquiries.
  Designed middle-ware for low-latency, concurrent request handling across multiple tenants.

Human Handoff System Jun 2024 — Nov 2024

• Objective: Facilitate seamless transitions from chatbot interactions to human agents for enhanced customer support.
• Details: Built a real-time chattting system using Node.js and Socket.IO with a queuing system for effective session management. Designed a fallback mechanism for agents to delegate interactions back to the chatbot when required. Preserved context during transitions to maintain conversational continuity and improve customer satisfaction.

Social Media Content Automation Software Jun 2024 — Nov 2024

• Objective: Automate the creation of social media posts for marketing managers using multi-agent systems.
• Details: Orchestrated multi-agent systems with LangChain, LangGraph, and LlamaIndex for task-specific operations like FAQs, captions, and comics generation. Leveraged GPT-4 and diffusion models for generating high-quality content. Integrated logging and feedback systems using LangSmith to refine agent performance over time. Combined libraries like BeautifulSoup and Pillow for web scraping and image manipulation, ensuring comprehensive content creation workflows.

Hand Orientation Detection Feb 2024 — Apr 2024

Utilized YOLOv8 and Python OpenCV to develop and train a custom model for  detecting and classifying diverse hand orientations. This involved gathering and  annotating a comprehensive dataset containing various hand orientations. The  YOLOv8 model was then trained using this dataset to accurately identify and  differentiate between different hand poses in real-time applications.

Reaction Timer Aug 2023 — Dec 2023

A VHDL-based system that is designed to replicate and measure human response to  stimuli over time. This digital setup integrates components such as a BCD counter,  a 7-segment display, a BCD decoder, flip-flops, and assorted logic gates. These  elements collaborate to generate stimuli and capture the user's reaction time to  them.

Purchase Order Management System Feb 2023 — May 2023

Management System, developed using C++, to automate the process of product  tracking and requisition. The system facilitates the process of creating purchase  requisitions and orders. and can generate reports of the current stock as well as the  recent requisitions.

Hydroponic Farming System Feb 2022 — May 2022

A hydroponic system, developed using C programming, mimics the functionality  of a real hydroponic setup by simulating the monitoring of various environmental  factors such as temperature and humidity. The system actively senses these  parameters and responds accordingly by either adjusting environmental conditions  or notifying the user of any deviations.