Usman G.

Key Skills

Software

Scale AI

Other

Top Subject Matter

Autonomous systems dataset labeling and RLHF/LLM alignment

Engineering research datasets for predictive maintenance ML

Legal Services & Contract Review

Top Data Types

Image

Document

Text

Top Task Types

Segmentation

Data Collection

Freelancer Overview

Data Annotation Specialist & AI Trainer | Scale AI / Outlier AI (Remote / Contract). Brings 2+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Scale AI and Other. Education includes Bachelor of Engineering, Ahmadu Bello University (2024). AI-training focus includes data types such as Image and Document and labeling workflows including Segmentation and Data Collection.

Labeling Experience

Data Annotation Specialist & AI Trainer | Scale AI / Outlier AI (Remote / Contract)

Scale AIImageSegmentation

Served as a Data Annotation Specialist and AI Trainer producing high-precision labels for autonomous systems and NLP model training. Annotated complex visual assets including bounding boxes and polygon/segmentation labels while also evaluating and rewriting generative outputs for RLHF alignment. Sustained a high quality bar by benchmarking accuracy and enforcing strict safety, helpfulness, and accuracy guidelines. • Labeled 10,000+ complex visual assets and corresponding text datasets monthly • Performed RLHF evaluation/rating and rewrite tasks to reduce hallucinations • Audited annotations against multi-page taxonomy manuals and edge-case guides • Collaborated with ML/data science teams to refine guides and reduce pipeline latency

2025 - Present

Graduate Technical Assistant | Ahmadu Bello University (ABU), Zaria

OtherDocumentData Collection

Worked as a Graduate Technical Assistant supporting research data preparation for engineering ML workflows. Managed and structured laboratory datasets by cleaning, normalizing, and formatting sensory and server-log data to make it suitable for training predictive maintenance models. Produced and maintained version-controlled documentation for research databases and institutional records. • Used Python (Pandas) to clean and normalize messy sensory and server-log data • Labeled hardware component metrics and server performance parameters for ML training • Authored and maintained version-controlled dataset documentation across shared folders • Supported research infrastructure for predictive maintenance algorithms

2023 - 2024

Education

Work History