Open to opportunities

Lubna Abbasi

@lubnaabbasi

Message

I build computer vision and NLP solutions for real-world applications.

Pakistan

Message

What I'm looking for

I'm looking for opportunities to build practical AI and machine learning solutions, especially in computer vision and NLP, while growing through real-world development, deployment, and collaborative engineering work.

I've developed machine learning and computer vision solutions through projects in emotion detection, image classification, object detection, and 3D crime-scene reconstruction.

At DAM TECHHUB, I developed and fine-tuned ResNet CNN models for FER2013 emotion detection, using augmentation and evaluation techniques to improve accuracy and F1-score. I also helped integrate real-time emotion recognition into retail-based applications.

For my final-year project at NUML, I'm building a 3D crime-scene reconstruction pipeline with COLMAP, spaCy, BERT, FastAPI, Unity, and YOLO. The system turns overlapping scene images and police reports into interactive visualizations, case summaries, and extracted evidence details.

Alongside AI work, I support IT service delivery at Vintage Managed Services through ticket coordination, troubleshooting, equipment inventory, vendor coordination, and procedure documentation.

Experience

Work history, roles, and key accomplishments

Current

IT Support Coordinator

Current

Vintage Managed Services

Sep 2025 - Present (11 months)

Assisted in coordinating daily IT support activities, including ticket handling, issue tracking and escalation. Supported end-users with basic troubleshooting for hardware, software and network issues.

IT Support Troubleshooting Hardware Software Networking Ticketing Systems Inventory Management Documentation

FYP 3D Crime Scene Construction Using NLP and Photogrammetry

National University of Modern Languages

Nov 2024 - Dec 2025 (1 year 1 month)

Built a 3D reconstruction pipeline using photogrammetry (COLMAP) for crime scenes based on overlapping images. Integrated NLP techniques (spaCy/BERT) for generating case summaries and extracting key entities from police reports.

Photogrammetry COLMAP NLP SpaCy BERT fastAPI Unity YOLO 3D Reconstruction Object Detection

Machine Learning Ai/ML Intern

DAM TECHHUB

Jul 2025 - Sep 2025 (2 months)

Developed and fine-tuned CNN models (ResNet) for emotion detection using FER2013 dataset. Applied data augmentation and evaluation techniques to improve model accuracy and F1-score.

Machine Learning Deep Learning CNN ResNet Model Evaluation Real Time Recognition