Alaa Aldin

Key Skills

Software

Anno-Mage

Clickworker

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Document

Top Task Types

Audio Recording

Bounding Box

Computer Programming Coding

Text Generation

Freelancer Overview

"A rigorous and methodical Data Quality Analyst with a background in Biomedical Engineering, transitioning technical analysis skills into the AI training domain. My expertise, developed over the past year, lies in meticulously validating and structuring diverse datasets. I specialize in evaluating the performance and safety of conversational AI models (LLMs), identifying biases, and ensuring factual consistency in generated responses. I have practical experience in both Arabic and English text annotation, encompassing advanced tasks such as semantic segmentation and complex categorization. Committed to maintaining extreme precision (QA/QC) and proficient in swiftly mastering new annotation platforms and project workflows."

Labeling Experience

LLM Quality Assurance and Safety Evaluation

ClickworkerTextEvaluation Rating

Involved in a continuous, high-volume project focused on enhancing the safety and factual grounding of a large language model (LLM). My core responsibility was performing prompt-response evaluation and side-by-side comparative rating of model outputs, strictly adhering to complex ethical and linguistic guidelines. Tasks included identifying and rating model hallucinations, bias, and potential safety risks (Red Teaming concept). I also contributed to prompt and response refinement (SFT) to improve Arabic and English conversational flow and accuracy. The work required an analytical, quality-control mindset to maintain an output accuracy rate exceeding 95%."

Involved in a continuous, high-volume project focused on enhancing the safety and factual grounding of a large language model (LLM). My core responsibility was performing prompt-response evaluation and side-by-side comparative rating of model outputs, strictly adhering to complex ethical and linguistic guidelines. Tasks included identifying and rating model hallucinations, bias, and potential safety risks (Red Teaming concept). I also contributed to prompt and response refinement (SFT) to improve Arabic and English conversational flow and accuracy. The work required an analytical, quality-control mindset to maintain an output accuracy rate exceeding 95%."

2024 - 2025

Education

Work History