Shahad Pk

Key Skills

Software

AWS SageMaker

Appen

Google Cloud Vertex AI

Label Studio

OpenCV AI Kit (OAK)

Roboflow

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Top Task Types

Bounding Box

Classification

Computer Programming Coding

Entity Ner Classification

Text Summarization

Freelancer Overview

I am an AI Trainer and Data Labeling Specialist with hands-on experience in text, image, and document annotation from both academic and freelance projects. During my AI internship, I worked on real-world systems including face recognition, document classification, and legal document matching, where I contributed to dataset creation, labeling, cleaning, and validation for model training. I am experienced with tools like Label Studio and cloud-based platforms, and I bring strong Python fundamentals with a practical understanding of NLP and computer vision workflows. I focus on producing clean, consistent training data that directly improves model performance in production environments.

Labeling Experience

Audio Transcription & Text Data Labeling Project

Label StudioTextEntity Ner ClassificationClassification

Worked on audio-to-text data preparation for AI training by first transcribing raw audio files into clean text format and then performing structured labeling using Label Studio. Tasks included text classification, entity tagging, and content segmentation to prepare high-quality datasets for downstream NLP model training. Also performed data cleaning and validation to ensure transcription accuracy, consistency, and usability for speech and language models.

Worked on audio-to-text data preparation for AI training by first transcribing raw audio files into clean text format and then performing structured labeling using Label Studio. Tasks included text classification, entity tagging, and content segmentation to prepare high-quality datasets for downstream NLP model training. Also performed data cleaning and validation to ensure transcription accuracy, consistency, and usability for speech and language models.

2025

Face Recognition Dataset Labeling for CCTV Surveillance

RoboflowImageBounding BoxClassification

Labeled face image data for a real-time CCTV-based face recognition system during my AI internship. Performed image annotation using bounding boxes and identity/class labels to build a clean training dataset. Supported preprocessing, dataset validation, and testing of recognition accuracy in real-world lighting and camera conditions. The labeled data was used to train and evaluate a deep learning–based face recognition model for attendance and security use cases.

Labeled face image data for a real-time CCTV-based face recognition system during my AI internship. Performed image annotation using bounding boxes and identity/class labels to build a clean training dataset. Supported preprocessing, dataset validation, and testing of recognition accuracy in real-world lighting and camera conditions. The labeled data was used to train and evaluate a deep learning–based face recognition model for attendance and security use cases.

2025

Legal Document Classification & Matching System

Label StudioImageBounding BoxEntity Ner Classification

Worked on a real-world legal document automation system during my AI internship. Performed large-scale text and document labeling using Label Studio for court records including petitions, orders, and case files. Tasks included document classification, entity tagging, and matching related legal documents for downstream model training using Layout-aware NLP pipelines. Also supported data cleaning, validation, and analysis of model predictions to improve accuracy and consistency of the document understanding system.

Worked on a real-world legal document automation system during my AI internship. Performed large-scale text and document labeling using Label Studio for court records including petitions, orders, and case files. Tasks included document classification, entity tagging, and matching related legal documents for downstream model training using Layout-aware NLP pipelines. Also supported data cleaning, validation, and analysis of model predictions to improve accuracy and consistency of the document understanding system.

2025

Education

Work History