Umar U.

Key Skills

Software

Other

Micro1

Mercor

Mindrift

Toloka

OneForma

Data Annotation Tech

Remotasks

Scale AI

Internal/Proprietary Tooling

OpenCV AI Kit (OAK)

Appen

AWS SageMaker

Clickworker

CloudFactory

CrowdFlower

CrowdSource

Anno-Mage

Argilla

Axiom AI

CVAT

Dataloop

Google Cloud Vertex AI

Label Studio

LightTag

LabelImg

Lionbridge

Mighty AI

Playment

Redbrick AI

Supervisely

SuperAnnotate

Snorkel AI

Sloth

Sama

Roboflow

VoTT

Trilldata Technologies

Telus

Top Subject Matter

Multilingual AI training and LLM output quality evaluation (Hausa/English)/ LUEL AI

Hausa speech data transcription and audio recording for AI training/ Mindrift AI

Subtitle generation and video annotation for multilingual AI conten/ Outlier AI

Top Data Types

Text

Audio

Image

Top Task Types

Transcription

Text Generation

Classification

Question Answering

Text Summarization

Object Detection

Computer Programming/Coding

Data Collection

Prompt + Response Writing (SFT)

Evaluation/Rating

RLHF

Function Calling

Freelancer Overview

I have hands-on experience contributing to AI training and evaluation projects through remote platforms such as Luel AI, Mindrift AI, and Outlier AI. In these roles, I worked on a variety of AI training data tasks including data labeling, prompt evaluation, response ranking, transcription, translation, and content annotation. I regularly evaluated AI-generated responses for instruction-following, factual accuracy, reasoning quality, formatting consistency, tone, and safety compliance. I also contributed to multilingual AI development by completing Hausa–English translation and transcription tasks, audio recordings for speech models, subtitle generation, and video annotation. These projects required careful adherence to detailed guidelines, strong analytical judgment, and consistent quality control when reviewing large volumes of AI-generated content. My background in Software Engineering and my Python Programming certification (88.94%) strengthen my ability to understand how training data influences machine learning models and AI behavior. I have also completed Python-based coding and problem-solving tasks, identifying issues such as hallucinations, weak reasoning, or policy violations in AI outputs. Combined with strong documentation and data organization skills using tools like Excel and collaborative remote workflows, I maintain a high standard of accuracy and efficiency. My multilingual ability in Hausa and English, fast adaptation to evolving project guidelines, and strong attention to detail allow me to contribute effectively to AI training, RLHF workflows, and data annotation projects.

Labeling Experience

AI Training & Evaluation Specialist — Subtitle generation/video annotation (Freelance / Remote)

OtherVideoText Generation

Worked on subtitle generation and video annotation tasks, reviewing timing and language accuracy. Ensured readability and content quality of generated subtitles for each segment. Applied structured checks to maintain consistency with annotation guidelines. • Subtitle generation with timing and readability review • Video annotation for language accuracy and content quality • Review for segment-level correctness and consistency • Structured evaluation against project guidelines

2023 - Present

AI Training & Evaluation Specialist — Hausa transcription/audio recording tasks (Freelance / Remote)

OtherAudioTranscription

Completed Hausa transcription and audio recording related tasks supporting AI speech and language model training workflows. Produced time-aligned language transcriptions and ensured content quality for downstream model use. Applied accuracy checks to maintain correctness of spoken-to-text outputs. • Hausa transcription tasks for speech/language model training • Audio recording support for AI training data • Quality verification of transcribed language content • Context preservation for multilingual training data

2023 - Present

AI Training & Evaluation Specialist — Freelance / Remote (AI training, labeling, and response evaluation)

OtherText

Performed AI training, evaluation, and quality review tasks across remote platforms including Luel AI, Mindrift AI, and Outlier AI. Evaluated AI-generated responses for instruction-following, factual accuracy, reasoning quality, formatting consistency, tone, and safety/policy compliance. Ranked and compared multiple AI outputs using project-specific rubrics to identify hallucinations, weak reasoning, and guideline violations. • Hausa↔English prompt and text translation while preserving context and accuracy • Ranked multiple responses to choose best outputs per rubric • Verified safety compliance, tone consistency, and formatting quality • Generalist AI data labeling, research, and typing-intensive content review

2023 - Present

Education

Work History