Ibrahim A.

Key Skills

Software

Scale AI

Toloka

Top Subject Matter

Large language model (LLM) output evaluation and RLHF training

AI content evaluation for labeling/QA and prompt-response improvement

Technical and Research Writing

Top Data Types

Document

Image

Video

Top Task Types

RLHF

Data Collection

Transcription

Text Summarization

Freelancer Overview

AI Data Specialist (Freelance) — Outlier.ai. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include research, outlier.ai and Luel. Education includes Bachelor of Pharmacy, University of Ilorin (2022). AI-training focus includes data types such as Text and labeling workflows including Evaluation and Rating.

Labeling Experience

AI Data Contributor (Freelance) — Luel

TextRLHF

Freelance AI Data Contributor role performing quality review and evaluation of AI-generated content against task instructions. Work emphasized identifying subtle differences between model outputs, flagging inconsistencies, and applying prompt engineering to improve response framing. Outputs supported downstream data quality assurance for AI training. • Reviewed AI content for correctness, coherence, and instruction alignment • Flagged errors and inconsistencies between closely related model outputs • Applied prompt engineering to test model responses and refine task framing • Performed rubric-like evaluation aligned to project guidelines

2026 - 2026

AI Data Specialist (Freelance) — Outlier.ai

TextData Collection

Freelance AI Data Specialist work focused on annotating and labeling LLM training data using detailed rubrics. The role involved evaluating AI-generated responses for accuracy, relevance, and overall quality, including hallucination and omission detection. Results were produced through structured review and preference judgment workflows to improve model behavior. • Rated and ranked model responses for RLHF preference judgments • Wrote and evaluated prompts for reasoning, instruction-following, and edge-case handling • Reviewed outputs for factual errors, hallucinations, and missing information; rewrote for clarity • Maintained consistent quality in high-volume repetitive review sessions

2026 - 2026

Education

Work History