Okas O.

Key Skills

Software

Don't disclose

Top Subject Matter

AI training for education and STEM LLM behavior evaluation

Rlhf Domain Expertise

and safety testing

Top Data Types

Text

Computer Code Programming

Top Task Types

RLHF

Prompt + Response Writing (SFT)

Freelancer Overview

Generalist AI Trainer – Stellar AI (Apr 2025 – Present). Brings 5+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose. Education includes Master of Arts, University of Roehampton and Bachelor of Education, University of Benin. AI-training focus includes data types such as Text and labeling workflows including RLHF and Prompt + Response Writing (SFT).

Labeling Experience

Generalist AI Trainer – Stellar AI (Apr 2025 – Present)

Don't discloseTextRLHF

Delivered high-accuracy annotation, evaluation, and prompt engineering for STEM, education, and general AI use cases. Assessed LLM outputs for accuracy, coherence, and logical consistency to improve reliability, while identifying hallucinations and weak reasoning patterns. Supported RLHF workflow improvements by refining training datasets and contributing to safety testing. • Produced and refined JSON-based training/evaluation datasets for LLMs • Performed structured output validation and rubric-driven evaluation • Conducted red-teaming/safety testing to identify vulnerabilities • Collaborated with cross-functional teams to improve annotation efficiency

2025 - Present

AI Tutor – Mindrift AI (Mar 2025 – Aug 2025)

Don't discloseTextPrompt + Response Writing (SFT)

Developed advanced prompts simulating real-world behavioural and educational scenarios. Evaluated AI-generated responses for alignment with expert-level knowledge and iteratively improved outputs through structured feedback loops. • Authored prompt scenarios for tutoring and behavior/education contexts • Reviewed and rated model responses for knowledge alignment • Created structured feedback for adaptive learning improvements • Supported prompt iteration based on evaluation findings

2025 - 2025

Education

Work History