Muhammad Umer Haroon

Academic

Career

Verbatim text

Muhammad Umer Haroon
Cell: 923365216427 |  Email: umer.haroon890@gmail.com
LinkedIn: https://www.linkedin.com/in/umerharoon
Address: G-10/4, MAIN SAWAN ROAD, HOUSE NO#431 , Islamabad , Pakistan
PROFESSIONAL PROFILE
I am an ambitious final year Electrical Engineering student at NUST with a strong background in Computer Vision, AI, robotics
systems, and digital/embedded design. I have worked on a range of academic and internship projects including vision-based
pipelines, machine learning models, and a ROS2-based autonomous robotic system for my Final Year Project. I have a solid
foundation in circuit design, microcontrollers, and system-level engineering, and I enjoy combining software, algorithms, and
hardware concepts to build practical, real-world systems. I bring strong problem-solving, teamwork, and leadership skills, with
experience in technical documentation, project coordination, and delivering complex systems under academic deadlines.
EDUCATION
Electrical Engineering
SEECS , Islamabad , 3.15 (2026)
INTERNSHIP EXPERIENCE
Systems Limited
10-Jun-2025 - 07-Aug-2025
During my AI/Data Science internship, I worked extensively on data preprocessing, cleaning, and exploratory data analysis (EDA)
using Python, Pandas, NumPy, Matplotlib, and Seaborn. I developed and evaluated machine learning models for classification tasks
and worked with real-world datasets from Kaggle and other sources. I fine-tuned BERT-based models for NLP tasks including spam
classification and named entity recognition, and implemented model evaluation and explainability using techniques such as SHAP
and LIME. I also worked with sentence embeddings (SBERT) for semantic similarity and clustering, gaining hands-on experience in
building end-to-end data-driven and AI pipelines using Jupyter Notebook and VS Code.
FINAL YEAR PROJECT
Lidar-Guided Rover for Landmine Detection
My Final Year Project is a ROS2-based LiDAR-guided autonomous ground rover designed for navigation in hazardous and disaster affected environments. The system integrates 2D LiDAR, SLAM, the Nav2 navigation stack, costmaps, and path planning to perform
autonomous mapping, localization, and obstacle avoidance in unknown environments. The rover is developed and tested in Gazebo
and visualized in RViz using a modular perception and navigation architecture that supports future sensor integration, including metal
detectors and thermal cameras for landmine detection and human presence detection. The project emphasizes safe and reliable
autonomous navigation, scalable system design, and real world applicability in high risk areas.
TECHNICAL EXPERTISE
Computer Vision & Image Processing: OpenCV, Feature Matching, Optical Flow, Homography
Experience building vision pipelines for motion analysis, object tracking, feature extraction, and geometric transformations using
Python and OpenCV.
Machine Learning & Deep Learning: scikit-learn, PyTorch, HuggingFace Transformers
Experience in training, fine-tuning, and evaluating machine learning and deep learning models for classification and NLP tasks using
real-world datasets.
Robotics & Autonomous Systems: ROS2, SLAM, Nav2, Gazebo, RViz
Experience developing and simulating autonomous navigation systems using LiDAR-based SLAM, path planning, localization, and

AI enrichment