Irtaqa Naveed

Key Skills

Software

Label Studio

Top Subject Matter

Enterprise HRMS

Large Language Models (LLMs)

Conversational AI

Top Data Types

Text

Video

Document

Audio

Image

Top Task Types

Object Detection

Classification

Fine-tuning

Transcription

Freelancer Overview

Senior AI Engineer — One Network, Islamabad. Brings 8+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Label Studio, Internal, and Proprietary Tooling. Education includes Master of Science, NUST (2025) and Bachelor of Science, University of Engineering and Technology, Taxila (2017). AI-training focus includes data types such as Text, Video, and Document and labeling workflows including Evaluation, Rating, and Object Detection.

Labeling Experience

Senior AI Engineer — One Network, Islamabad

Label StudioText

Led structured output validation of large language model (LLM) responses for enterprise and HRMS domains. Designed and implemented annotation standards for tool-call accuracy and response grounding. Built frameworks to detect hallucinations, factual inconsistencies, and instruction-following errors. • Created and enforced labeling schemas for preference ranking, NER, and intent classification. • Annotated SQL correctness and RLHF preference pairs for fine-tuning. • Verified and validated agentic behavior across multi-step reasoning chains. • Utilized Label Studio and internal pipelines for structured annotation.

Led structured output validation of large language model (LLM) responses for enterprise and HRMS domains. Designed and implemented annotation standards for tool-call accuracy and response grounding. Built frameworks to detect hallucinations, factual inconsistencies, and instruction-following errors. • Created and enforced labeling schemas for preference ranking, NER, and intent classification. • Annotated SQL correctness and RLHF preference pairs for fine-tuning. • Verified and validated agentic behavior across multi-step reasoning chains. • Utilized Label Studio and internal pipelines for structured annotation.

2025 - 2025

Machine Learning Engineer — Rapidev, Islamabad

VideoObject Detection

Curated a 15k-frame annotated video dataset for YOLOv8 object detection model training. Led annotation of bounding boxes, object class labels, occlusion tags, and edge-case samples. Established strict standards for facial recognition and traffic violation detection datasets. • Developed annotation protocols for bounding boxes and object classification. • Tagged occlusions and handled edge-case identification in video frames. • Standardized data for consistency and future retraining. • Used internal proprietary tooling tailored for computer vision workflows.

Curated a 15k-frame annotated video dataset for YOLOv8 object detection model training. Led annotation of bounding boxes, object class labels, occlusion tags, and edge-case samples. Established strict standards for facial recognition and traffic violation detection datasets. • Developed annotation protocols for bounding boxes and object classification. • Tagged occlusions and handled edge-case identification in video frames. • Standardized data for consistency and future retraining. • Used internal proprietary tooling tailored for computer vision workflows.

2024 - 2025

Senior Systems Lead (IT Branch) — Pakistan Air Force

DocumentClassification

Managed end-to-end data annotation and labeling workflows for defense-sector automated decision-support systems. Led the structuring, labeling, and quality assurance of mission-critical document and workflow datasets. Maintained high standards for labeling accuracy, data compliance, and integrity in sensitive applications. • Coordinated structured data labeling for automated workflows and AI models. • Enforced data compliance protocols across multiple annotation pipelines. • Oversaw quality assurance and inter-annotator agreement enforcement. • Utilized secure internal platforms designed for classified data environments.

Managed end-to-end data annotation and labeling workflows for defense-sector automated decision-support systems. Led the structuring, labeling, and quality assurance of mission-critical document and workflow datasets. Maintained high standards for labeling accuracy, data compliance, and integrity in sensitive applications. • Coordinated structured data labeling for automated workflows and AI models. • Enforced data compliance protocols across multiple annotation pipelines. • Oversaw quality assurance and inter-annotator agreement enforcement. • Utilized secure internal platforms designed for classified data environments.

2017 - 2024

Multilingual Audio Annotation

Label StudioAudioTranscription

Annotated multilingual audio datasets for transcription quality, translation accuracy, and denoising evaluations in low-SNR environments. Ensured accurate tagging of language content and cross-validated translation standards. Supported the improvement of speech models for noisy operational settings. • Evaluated transcription performance across diverse languages. • Flagged translation errors and reviewed denoising effectiveness. • Utilized structured annotation schemas for audio data labeling. • Performed quality assessment with Label Studio and internal tools.

Annotated multilingual audio datasets for transcription quality, translation accuracy, and denoising evaluations in low-SNR environments. Ensured accurate tagging of language content and cross-validated translation standards. Supported the improvement of speech models for noisy operational settings. • Evaluated transcription performance across diverse languages. • Flagged translation errors and reviewed denoising effectiveness. • Utilized structured annotation schemas for audio data labeling. • Performed quality assessment with Label Studio and internal tools.

Not specified

LLM Fine-Tuning Data Pipeline (LLaMA 3.2)

Label StudioTextFine Tuning

Labeled and curated instruction-response pairs, preference pairs, tool-call annotations, and RLHF ranking data for LLM fine-tuning. Focused on datasets for HRMS and enterprise applications, supporting high-quality model improvements. Ensured consistent adherence to provided schemas, formats, and annotation guidelines. • Managed RLHF pairwise preference and structured output validation. • Executed multi-step tool-call annotation and scoring for agentic evaluations. • Performed instruction-following and factuality reviews for sample responses. • Used structured schemas and JSONL formats via Label Studio and custom pipelines.

Labeled and curated instruction-response pairs, preference pairs, tool-call annotations, and RLHF ranking data for LLM fine-tuning. Focused on datasets for HRMS and enterprise applications, supporting high-quality model improvements. Ensured consistent adherence to provided schemas, formats, and annotation guidelines. • Managed RLHF pairwise preference and structured output validation. • Executed multi-step tool-call annotation and scoring for agentic evaluations. • Performed instruction-following and factuality reviews for sample responses. • Used structured schemas and JSONL formats via Label Studio and custom pipelines.

Not specified

Education

Work History