Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I’m Noah Jatuwase Tolulope, an AI Trainer and RLHF specialist based in Lagos, Nigeria. For the past four years, I’ve worked directly on large language model training pipelines across OpenAI and Anthropic platforms—building high-quality instruction-response pairs, ranking model outputs, and stress-testing prompts to uncover real-world failure modes. I’m also a prompt engineer who enjoys making systems more reliable through structured rubrics, careful QA, and repeatable templates that improve both output quality and annotation throughput. With a background in Information Systems from FUTA, I approach labeling and evaluation like a data pipeline problem—so the work stays consistent at scale while maintaining strong quality scores.…I’m Noah Jatuwase Tolulope, an AI Trainer and RLHF specialist based in Lagos, Nigeria. For the past four years, I’ve worked directly on large language model training pipelines across OpenAI and Anthropic platforms—building high-quality instruction-response pairs, ranking model outputs, and stress-testing prompts to uncover real-world failure modes. I’m also a prompt engineer who enjoys making systems more reliable through structured rubrics, careful QA, and repeatable templates that improve both output quality and annotation throughput. With a background in Information Systems from FUTA, I approach labeling and evaluation like a data pipeline problem—so the work stays consistent at scale while maintaining strong quality scores.

Noah Jatuwase Tolulope

AI Trainer, AI Prompt Engineer, AI Evaluator, +2





I’m Noah Jatuwase Tolulope, an AI Trainer and RLHF specialist based in Lagos, Nigeria. For the past four years, I’ve worked directly on large language model training pipelines across OpenAI and Anthropic platforms—building high-quality instruction-response pairs, ranking model outputs, and stress-testing prompts to uncover real-world failure modes. I’m also a prompt engineer who enjoys making systems more reliable through structured rubrics, careful QA, and repeatable templates that improve both output quality and annotation throughput. With a background in Information Systems from FUTA, I approach labeling and evaluation like a data pipeline problem—so the work stays consistent at scale while maintaining strong quality scores.…I’m Noah Jatuwase Tolulope, an AI Trainer and RLHF specialist based in Lagos, Nigeria. For the past four years, I’ve worked directly on large language model training pipelines across OpenAI and Anthropic platforms—building high-quality instruction-response pairs, ranking model outputs, and stress-testing prompts to uncover real-world failure modes. I’m also a prompt engineer who enjoys making systems more reliable through structured rubrics, careful QA, and repeatable templates that improve both output quality and annotation throughput. With a background in Information Systems from FUTA, I approach labeling and evaluation like a data pipeline problem—so the work stays consistent at scale while maintaining strong quality scores.

Available to hire

I’m Noah Jatuwase Tolulope, an AI Trainer and RLHF specialist based in Lagos, Nigeria. For the past four years, I’ve worked directly on large language model training pipelines across OpenAI and Anthropic platforms—building high-quality instruction-response pairs, ranking model outputs, and stress-testing prompts to uncover real-world failure modes.

I’m also a prompt engineer who enjoys making systems more reliable through structured rubrics, careful QA, and repeatable templates that improve both output quality and annotation throughput. With a background in Information Systems from FUTA, I approach labeling and evaluation like a data pipeline problem—so the work stays consistent at scale while maintaining strong quality scores.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Language

English

Fluent

Work Experience

AI Trainer and RLHF Specialist at Freelance / Remote Contracts

January 1, 2022 - Present

Reviewed and ranked over 12,000 model responses across OpenAI and Anthropic training platforms, consistently maintaining high accuracy (above 95%). Wrote more than 1,500 instruction-response pairs for supervised fine-tuning datasets spanning reasoning, code explanations, creative writing, and factual QA. Built and ran 300+ adversarial prompt sequences to probe safety boundaries and fed results into policy-aligned behavior updates. Created personal annotation rubrics to reduce ambiguity, improving throughput by about 30% without sacrificing quality. Performed pairwise preference ranking for multi-turn conversations and documented rationales to help training teams understand why responses were preferred. Delivered batches of 500–1,000 tasks on schedule with no quality escalations.

Prompt Engineer at Independent Projects

January 1, 2021 - December 31, 2022

Developed a reusable library of 200+ structured prompts for GPT-3, GPT-4, and Claude tailored to tasks like summarization, classification, and data extraction. Reduced prompt iteration cycles from 8 rounds to 3 by logging failure modes systematically, saving about 40% of testing time. Applied chain-of-thought, few-shot, and role-based prompting to improve output quality on complex reasoning tasks by an estimated 25–35% versus zero-shot baselines. Produced prompt templates with annotated examples and edge-case notes for reuse across multiple projects.