Open to opportunities

Phillip Afa

@phillipafa

Message

Detail-obsessed AI training specialist delivering high-accuracy data and LLM evaluations.

Nigeria

Message

What I'm looking for

I’m looking for remote AI training work where quality and consistency matter—LLM evaluation, RLHF-adjacent feedback, and multi-modal annotation. I can deliver high-throughput labeling and clear, guideline-faithful outputs for production datasets.

I’m a detail-obsessed AI training specialist focused on high-throughput, quality-first data annotation for modern AI systems. Over 6+ years, I’ve delivered 43,000+ tasks with industry-leading accuracy, including 96.5% peak approval accuracy on MTurk.

My work spans multi-modal annotation—text, image, audio, and video—plus structured LLM evaluation and AI response rating/ranking. I contribute to RLHF-adjacent feedback pipelines by producing structured human preference signals that support model training, refinement, and quality benchmarking.

I bring domain depth from medical sciences, biological sciences, creative writing, and marketing, which helps me label and evaluate with consistency and context. I’m available remotely for AI training companies seeking reliable annotation professionals who can maintain accuracy across large-scale batches.

Experience

Work history, roles, and key accomplishments

Current

Annotator & Image Labeling

Current

Toloka AI (Yandex)

Jan 2023 - Present (3 years 3 months)

Supported computer vision image classification and labeling projects while completing ~94 tasks with ~92% accuracy. Also performed audio transcription and text annotation, applying judgment to edge cases to provide high-quality signals for AI training datasets.

Image Labeling Edge Cases Audio Transcription Annotating Data Quality Assurance Datasets

Current

Data Annotator

Current

Oneforma (Premise Data)

Jan 2023 - Present (3 years 3 months)

Contributed to ongoing text, image, and audio/multimodal annotation pipelines with ~92% accuracy across ~94 tasks by following project-specific guidelines. Worked on NER tagging, sentiment classification, audio transcription, and image labeling within structured QA frameworks to maintain annotation integrity.

Audio Transcription Image Labeling Data Quality Assurance

Current

HIT Specialist

Current

Amazon Mechanical Turk (MTurk)

Jan 2016 - Present (10 years 3 months)

Completed 28,000+ HITs with 96.5% approval accuracy, performing AI response rating and ranking to support LLM output evaluation. Also contributed audio transcription and AI image generation evaluation, providing structured preference signals for RLHF-adjacent feedback pipelines.

Data Annotation LLM Output Evaluation Audio Transcription RLHF Feedback Signals Prompt Alignment Checks Quality Assurance

Data Annotator

Clickworker / UHRS Platform

Jan 2019 - Jul 2020 (1 year 6 months)

Submitted and received approval for 15,000+ annotation tasks across diverse AI training datasets, including multi-modal labeling and classification. Performed bounding box annotation, NER, sentiment tagging, and audio transcription while rating and ranking AI-generated responses for RLHF data collection workflows.

Bounding Boxes Audio Transcription Task Guideline Compliance

Education

Degrees, certifications, and relevant coursework

University of Uyo

Bachelor of Science, Fisheries & Aquatic Management

Earned a Bachelor of Science in Fisheries & Aquatic Management. Relevant biological sciences training supports medical, veterinary, and scientific domain annotation work.