Raheel A.

Key Skills

Software

Don't disclose

Appen

Clickworker

Labelbox

Label Studio

Mercor

Other

Internal/Proprietary Tooling

Top Subject Matter

AI & Machine Learning - Appen, Shaip, Bhasha, ADAP — direct experience

Language & Linguistics - M.A. Sindhi, native multilingual, 5+ years

Education & E-Learning - Government teacher, curriculum experience

Top Data Types

Audio

Text

Image

Document

Top Task Types

Transcription

Bounding Box

Polygon

Segmentation

Classification

Point/Key Point

Polyline

Object Detection

Text Generation

Question Answering

Text Summarization

Evaluation/Rating

Data Collection

Fine-tuning

Freelancer Overview

I have over five years of hands-on experience contributing to AI training data projects as a native Sindhi and Urdu language specialist. My work spans verbatim speech transcription with timestamping and speaker labeling (ADAP), machine-generated text auditing and quality assurance (Bhasha Platform), multimodal annotation including voice, text, and image data (Appen/ByteDance), and Sindhi linguistic content validation for ML datasets (Shaip). I have consistently maintained accuracy benchmarks of 98%+ and am experienced with annotation schemas, style guide compliance, and error classification frameworks used across major AI data platforms. What sets me apart is my native command of Sindhi — a critically low-resource language that very few qualified AI data contributors can offer. Combined with native Urdu proficiency and advanced English, I cover three languages across formal, conversational, literary, legal, medical, and journalistic domains. My academic background — a Master's in Sindhi Language & Literature (First Division, University of Sindh) — further strengthens my linguistic judgment beyond what a non-specialist annotator can provide. I have worked with Google LOFT 2.0, Appen, Shaip, Bhasha, and ADAP, giving me cross-platform familiarity with the standards and workflows that enterprise AI data projects demand.

Labeling Experience

Urdu AI Speech Transcription Specialist (ADAP via Upwork)

Don't discloseAudioTranscription

Urdu audio transcription for AI training datasets at 98%+ accuracy. Included timestamping, multi-speaker labeling, and non-speech event annotation to support high-quality supervised learning. Ensured deliverables consistently followed annotation schemas and client style guidance. • Verbatim Urdu transcription • Timestamping and speaker labeling • Non-speech event annotation • Style guide and schema compliance

2025 - Present

Sindhi Language Subject Matter Expert (SME) (Shaip)

Don't discloseText

Reviewed and validated Sindhi linguistic content intended for AI/ML training datasets. Assessed fluency, grammar correctness, and cultural appropriateness of machine-generated outputs. Supplied expert feedback to strengthen low-resource Sindhi language data quality. • Linguistic fluency and grammar checks • Cultural appropriateness validation • Feedback to improve low-resource data • Dataset QA and expert review

2025 - 2025

Urdu Language SME & Content Auditor (Bhasha Platform)

Text

Performed Urdu language expert auditing of machine-generated sentences for AI training and content quality. Conducted error classification and correction across news, social media, and conversational domains. Maintained throughput while meeting strict accuracy benchmarks. • Linguistic quality and accuracy auditing • Error classification and correction • Domain coverage (news/social/conversation) • Throughput with accuracy benchmarks

2024 - 2025

Urdu / Sindhi Language Specialist & QA Analyst (Appen via ByteDance)

Don't discloseAudio

Supported multilingual AI training via translation, MT post-editing, and content annotation tasks. Reviewed and rated Urdu and Sindhi outputs for naturalness, fluency, and factual accuracy. Annotated voice, text, and image data to improve dataset coverage for multilingual models. • Translation and MT post-editing • Rating for naturalness and fluency • Factual accuracy verification • Voice/text/image data annotation

2023 - 2024

Sindhi Translator — TEA Project (Google Crowdsource / LOFT 2.0)

CrowdSourceText

Translated diverse English content into Sindhi for a crowdsourced TEA project. Covered legal notices, government documents, product descriptions, news articles, and general-purpose text. Ensured terminological accuracy for low-resource Sindhi language data destined for AI training corpora. • Legal and government document translation • News and product translation • Terminology accuracy for low-resource data • General-purpose text localization

2023 - 2023

Education

Work History