Ehsan Waqar

Key Skills

Software

Appen

Clickworker

Surge AI

Labelbox

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Document

Image

Top Task Types

Audio Recording

Data Collection

Evaluation Rating

Question Answering

Translation Localization

Freelancer Overview

I bring over three years of hands-on experience in data labeling, AI training data QA, and LLM output evaluation for leading global projects. My work has focused on improving the quality and relevance of AI assistant conversations through precise fact-checking, intent alignment, and prompt refinement. I’ve contributed to generative AI datasets by labeling question-answer pairs, reviewing conversational outputs, and providing structured feedback to enhance model performance and user alignment.

Labeling Experience

RWS

Data Annotation TechTextEvaluation Rating

Actively involved in the annotation and quality assurance of large language model (LLM) outputs for English conversational datasets. Tasks included rewriting user prompts, verifying assistant intent alignment, labeling question-answer pairs, and evaluating multi-turn dialog responses. Delivered consistent high-accuracy results (>95%) across QA reviews, NER tagging, and instruction-based prompt refinement. Communicated feedback in structured formats to help optimize LLM behavior for real-world use cases. Operated in a fully remote environment using version-controlled workflows and detailed guidelines.

Actively involved in the annotation and quality assurance of large language model (LLM) outputs for English conversational datasets. Tasks included rewriting user prompts, verifying assistant intent alignment, labeling question-answer pairs, and evaluating multi-turn dialog responses. Delivered consistent high-accuracy results (>95%) across QA reviews, NER tagging, and instruction-based prompt refinement. Communicated feedback in structured formats to help optimize LLM behavior for real-world use cases. Operated in a fully remote environment using version-controlled workflows and detailed guidelines.

2024

CrowdGen

LabelboxTextQuestion AnsweringText Summarization

Contributed to high-accuracy annotation and QA review of LLM-generated outputs for English-language conversational datasets. Tasks included prompt rewriting, validation of assistant replies, error flagging, and user intent alignment. Maintained >95% accuracy on reviewed tasks and provided structured feedback for model fine-tuning. Also supported instruction tuning and dialog dataset curation using remote QA workflows and version-controlled guidelines.

Contributed to high-accuracy annotation and QA review of LLM-generated outputs for English-language conversational datasets. Tasks included prompt rewriting, validation of assistant replies, error flagging, and user intent alignment. Maintained >95% accuracy on reviewed tasks and provided structured feedback for model fine-tuning. Also supported instruction tuning and dialog dataset curation using remote QA workflows and version-controlled guidelines.

2022 - 2025

Education

Work History