Linus V.

Key Skills

Software

Don't disclose

Other

Top Subject Matter

Multidomain AI evaluation (STEM + general knowledge)

Multimodal dataset annotation and QA

Programming and math task validation for AI training

Top Data Types

Text

Image

Video

Document

Top Task Types

Classification

RLHF

Freelancer Overview

Generalist AI Trainer & Evaluator (LLM evaluation and RLHF feedback) — Sahara AI,Mercor, Abaka AI. Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose, Internal, and Proprietary Tooling. Education includes Bachelor of Science, University of Ibadan (2026) and West African Senior School Certificate Examination, N/A (2019). AI-training focus includes data types such as Text, Image, and Computer Code and labeling workflows including Evaluation, Rating, and Classification.

Labeling Experience

Multimodal quality assessment — Sahara AI

Don't discloseVideo

Assessed AI-generated image and video content for quality, coherence, and alignment with task requirements. Produced human feedback reports that highlighted artifacts, semantic mismatches, and spatial inaccuracies. Used these assessments to support model training improvements. • Rated visual outputs for task alignment and overall coherence • Reported detected artifacts and content inconsistencies • Generated structured feedback for training/model teams • Supported iterative quality improvements through evaluation

2024 - Present

LLM evaluation feedback contributing to RLHF — Sahara AI

Don't discloseTextRLHF

Delivered human feedback on AI-generated outputs used directly in RLHF pipelines to improve model accuracy and safety metrics. Authored structured feedback reports that reflected rubric criteria and actionable guidance. Used error pattern analysis to inform model refinement priorities. • Provided rubric-aligned human ratings and written feedback • Supported RLHF feedback loops with structured scoring • Documented failure modes for subsequent fine-tuning guidance • Contributed to quality and safety metric improvements

2024 - Present

Generalist AI Trainer & Evaluator (LLM evaluation and RLHF feedback) — Sahara AI

Don't discloseText

Evaluated rubric-based LLM-generated responses across reasoning, coding, mathematics, physics, and general knowledge domains. Scored outputs for quality, coherence, instruction-following, and safety using structured criteria that supported RLHF workflows. Identified recurring failure patterns and produced detailed error reports to guide iterative model fine-tuning. • Built and applied multi-criteria evaluation rubrics for 500+ model responses • Conducted prompt analysis and iterative prompt refinement for better model behavior • Performed prompt + response quality assessment to rank and score candidate outputs • Produced structured feedback artifacts for RLHF pipeline consumption

2024 - Present

Data labeling support activities embedded within Data Annotator role — ABAKA AI

Other

Validated programming tasks and logical correctness during labeling and evaluation activities for AI training datasets. Contributed to dataset annotation where code correctness and clarity were key acceptance criteria. Ensured labeled items aligned with task requirements and internal quality standards. • Reviewed code/math solution correctness and completeness • Checked annotation quality for logical consistency across evaluator teams • Supported QA for training/evaluation dataset readiness • Flagged issues requiring rework to improve dataset quality

2024 - 2024

Data Annotator — ABAKA AI (Astra Bright Analytics & Knowledge Alliance)

ImageClassification

Annotated and labeled multimodal AI training data for text, image, audio, and video modalities as part of dataset preparation. Performed quality checks to ensure the annotated data met benchmark standards for downstream model evaluation and training. Validated mathematical solutions and programming tasks for logical accuracy, clarity, and completeness. • Labeled image-related training examples with attention to benchmark requirements • Supported model evaluation and quality assurance workflows • Ensured labeling consistency and correctness for math and programming components • Helped validate dataset outputs prior to use in evaluation pipelines

2024 - 2024

Education

Work History