Ishaq Aiyelabegan

Key Skills

Software

Labelbox

Top Subject Matter

Autonomous Vehicles / Computer Vision

Audio Transcription and NLP

Large Language Models (LLMs) / NLP

Top Data Types

Image

Audio

Text

Document

Top Task Types

Bounding Box

Transcription

RLHF

Freelancer Overview

Data Annotation Specialist & QA Lead. Brings 5+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Labelbox, Otter.ai, and Internal. Education includes Bachelor of Engineering, Ahmadu Bello University (2023). AI-training focus includes data types such as Image, Audio, and Text and labeling workflows including Bounding Box, Transcription, and RLHF.

Labeling Experience

Data Annotation Specialist & QA Lead

LabelboxImageBounding Box

Labeled over 7,500 images for autonomous-vehicle perception using bounding boxes, polygons, and keypoints. Achieved a high first-pass QA acceptance rate of 95% across four client datasets. Ensured annotation quality and consistency by leading calibration sessions and mentoring team members. • Applied advanced annotation techniques for self-driving datasets. • Maintained high productivity under tight deadlines. • Consistently adhered to strict project guidelines. • Utilized tools like Labelbox and SuperAnnotate for annotation tasks.

Labeled over 7,500 images for autonomous-vehicle perception using bounding boxes, polygons, and keypoints. Achieved a high first-pass QA acceptance rate of 95% across four client datasets. Ensured annotation quality and consistency by leading calibration sessions and mentoring team members. • Applied advanced annotation techniques for self-driving datasets. • Maintained high productivity under tight deadlines. • Consistently adhered to strict project guidelines. • Utilized tools like Labelbox and SuperAnnotate for annotation tasks.

2023 - Present

Transcription QA Reviewer & Annotation Contributor

AudioTranscription

Transcribed over 600 hours of multi-speaker audio including legal, medical, and financial content to clean-verbatim standards. Conducted QA reviews of 420 subtitle files for timing accuracy and compliance. Produced NER, sentiment, and intent labels for a conversational AI dataset following detailed taxonomy guidelines. • Achieved 98.1% word accuracy in transcription tasks. • Reviewed subtitles with 99.2% error-free output. • Generated 180,000 annotated tokens for AI language corpus. • Used Otter.ai and Aegisub for efficient audio and subtitle annotation.

Transcribed over 600 hours of multi-speaker audio including legal, medical, and financial content to clean-verbatim standards. Conducted QA reviews of 420 subtitle files for timing accuracy and compliance. Produced NER, sentiment, and intent labels for a conversational AI dataset following detailed taxonomy guidelines. • Achieved 98.1% word accuracy in transcription tasks. • Reviewed subtitles with 99.2% error-free output. • Generated 180,000 annotated tokens for AI language corpus. • Used Otter.ai and Aegisub for efficient audio and subtitle annotation.

2022 - 2023

AI Training Data Reviewer

TextRLHF

Completed RLHF ranking tasks on more than 4,500 response pairs to fine-tune language models. Authored over 300 adversarial prompts for red-teaming to evaluate model safety and robustness. Ensured outputs contributed to client safety testing and instruction-following evaluations. • Evaluated language model responses for accuracy and helpfulness. • Performed prompt engineering for adversarial testing. • Assessed tone and safety adherence in AI outputs. • Used Google Docs and Internal/Proprietary Tooling for annotation.

Completed RLHF ranking tasks on more than 4,500 response pairs to fine-tune language models. Authored over 300 adversarial prompts for red-teaming to evaluate model safety and robustness. Ensured outputs contributed to client safety testing and instruction-following evaluations. • Evaluated language model responses for accuracy and helpfulness. • Performed prompt engineering for adversarial testing. • Assessed tone and safety adherence in AI outputs. • Used Google Docs and Internal/Proprietary Tooling for annotation.

2022 - 2022

Education

Work History