Elijah Ogaba

Key Skills

Software

Other

CVAT

Top Subject Matter

General Computer Vision/Medical

Autonomous Systems/Video Analysis

Text Classification/NLP Safety

Top Data Types

Text

Document

Image

Video

Audio

Top Task Types

Text Summarization

Entity Ner Classification

Segmentation

Tracking

RLHF

Bounding Box

Action Recognition

Classification

Transcription

Freelancer Overview

Image Annotation Specialist (Bounding Box & Attribute Labelling). Core strengths include CVAT and Other. AI-training focus includes data types such as Image, Video, and Text and labeling workflows including Bounding Box, Action Recognition, and Classification.

Labeling Experience

LLM Output Comparative Evaluation — AI Model Assessor

Text

I conducted structured output comparisons of LLMs using diverse prompt architectures to evaluate scientific reasoning depth, factual accuracy, and summary quality. The workflow included side-by-side evaluations of outputs from Claude, ChatGPT, and Gemini based on identical research tasks in genomics and diagnostics. I documented specific improvements and discrepancies resulting from different annotation strategies such as Chain-of-Thought and zero-shot prompting. • Ran comparative LLM evaluations on genomics and clinical lab data tasks. • Measured structure, depth, and specificity improvements using annotated prompt experiments. • Provided quantitative and qualitative feedback on LLM performance in scientific and medical domains. • Generated detailed reports for multi-model assessment of scientific output quality.

I conducted structured output comparisons of LLMs using diverse prompt architectures to evaluate scientific reasoning depth, factual accuracy, and summary quality. The workflow included side-by-side evaluations of outputs from Claude, ChatGPT, and Gemini based on identical research tasks in genomics and diagnostics. I documented specific improvements and discrepancies resulting from different annotation strategies such as Chain-of-Thought and zero-shot prompting. • Ran comparative LLM evaluations on genomics and clinical lab data tasks. • Measured structure, depth, and specificity improvements using annotated prompt experiments. • Provided quantitative and qualitative feedback on LLM performance in scientific and medical domains. • Generated detailed reports for multi-model assessment of scientific output quality.

2025 - Present

AI Tools Specialist — Prompt Engineer

Text

As an AI Tools Specialist and Prompt Engineer, I designed and evaluated nineteen AI prompt engineering systems across Chain-of-Thought, Role Prompting, Few-Shot, and System Prompt architectures within scientific domains. My work involved evaluating LLM outputs for factual accuracy, coherence, logical reasoning, hallucination, and bias, including multi-step Chain-of-Thought prompts with self-checking. I also built systems for scientific claim evaluation, clinical trial summarization, and regulatory compliance guidance in health and biotech areas. • Led comparative LLM evaluations using AI models such as Claude, GPT-4, and Gemini. • Designed, annotated, and validated multi-step scientific reasoning chains across clinical and regulatory contexts. • Built end-to-end evaluation flows for scientific, investor, and clinical documentation tasks. • Maintained high data integrity standards and domain-specific annotation protocols throughout all case studies.

As an AI Tools Specialist and Prompt Engineer, I designed and evaluated nineteen AI prompt engineering systems across Chain-of-Thought, Role Prompting, Few-Shot, and System Prompt architectures within scientific domains. My work involved evaluating LLM outputs for factual accuracy, coherence, logical reasoning, hallucination, and bias, including multi-step Chain-of-Thought prompts with self-checking. I also built systems for scientific claim evaluation, clinical trial summarization, and regulatory compliance guidance in health and biotech areas. • Led comparative LLM evaluations using AI models such as Claude, GPT-4, and Gemini. • Designed, annotated, and validated multi-step scientific reasoning chains across clinical and regulatory contexts. • Built end-to-end evaluation flows for scientific, investor, and clinical documentation tasks. • Maintained high data integrity standards and domain-specific annotation protocols throughout all case studies.

2025 - Present

AI Tools Specialist & Domain Expert Annotator

CVATVideoTrackingAction Recognition

I executed high-precision temporal object tracking projects for diverse video datasets using CVAT. My work utilized interpolation techniques to maintain consistent object IDs across multiple frames in dynamic scenes, such as aerial drone and retail videos. I enriched metadata by applying attribute-level annotations to ensure fit-for-purpose datasets for machine learning models. • Applied linear and non-linear interpolation for temporal object consistency. • Categorized and annotated multi-class objects with complex attribute schemas. • Ensured high-quality exports for autonomous system and retail-focused AI workflows. • Conducted rigorous QA of bounding coordinates and attribute consistency.

I executed high-precision temporal object tracking projects for diverse video datasets using CVAT. My work utilized interpolation techniques to maintain consistent object IDs across multiple frames in dynamic scenes, such as aerial drone and retail videos. I enriched metadata by applying attribute-level annotations to ensure fit-for-purpose datasets for machine learning models. • Applied linear and non-linear interpolation for temporal object consistency. • Categorized and annotated multi-class objects with complex attribute schemas. • Ensured high-quality exports for autonomous system and retail-focused AI workflows. • Conducted rigorous QA of bounding coordinates and attribute consistency.

2024 - Present

AI Tools Specialist & Domain Expert Annotator

CVATImageSegmentationBounding Box

I executed high-precision instance segmentation projects for hematological datasets using CVAT and Labelbox. My work included annotating RBCs, WBCs, and platelets to a high level of anatomical accuracy, focusing on medical image data for AI training. I ensured data integrity and export quality for integration into machine learning pipelines. • Performed pixel-level polygon segmentation and object identification using CVAT and Labelbox. • Managed COCO JSON and YOLO file formats to guarantee compatibility with model training workflows. • Verified and reviewed exported data formats for quality assurance in annotation projects. • Applied clinical laboratory expertise for high-accuracy biological plausibility in label consistency.

I executed high-precision instance segmentation projects for hematological datasets using CVAT and Labelbox. My work included annotating RBCs, WBCs, and platelets to a high level of anatomical accuracy, focusing on medical image data for AI training. I ensured data integrity and export quality for integration into machine learning pipelines. • Performed pixel-level polygon segmentation and object identification using CVAT and Labelbox. • Managed COCO JSON and YOLO file formats to guarantee compatibility with model training workflows. • Verified and reviewed exported data formats for quality assurance in annotation projects. • Applied clinical laboratory expertise for high-accuracy biological plausibility in label consistency.

2024 - Present

Chain-of-Thought Evaluation Projects — Annotation Specialist

OtherDocumentText SummarizationEntity Ner Classification

In my current role as a Data Labelling Specialist and AI Annotator, I perform complex NER annotation and relation extraction on clinical medical texts. My responsibilities involve identifying and tagging a broad range of entity types and extracting relationships, completing coreference resolution, multi-dimension quality labeling, and hallucination detection for AI and LLM outputs. I specialize in edge case identification and calibration-standard annotation within the health and life sciences domain. • Tagged eleven clinical entity types such as person, drug, dosage, and disease in medical texts. • Applied nine clinical relation categories including prescriptive, causal, and temporal mappings. • Independently rated AI output quality along accuracy, helpfulness, instruction following, safety, and clarity. • Detected hallucinations, completed preference annotation tasks, and performed multi-step Chain-of-Thought labeling.

In my current role as a Data Labelling Specialist and AI Annotator, I perform complex NER annotation and relation extraction on clinical medical texts. My responsibilities involve identifying and tagging a broad range of entity types and extracting relationships, completing coreference resolution, multi-dimension quality labeling, and hallucination detection for AI and LLM outputs. I specialize in edge case identification and calibration-standard annotation within the health and life sciences domain. • Tagged eleven clinical entity types such as person, drug, dosage, and disease in medical texts. • Applied nine clinical relation categories including prescriptive, causal, and temporal mappings. • Independently rated AI output quality along accuracy, helpfulness, instruction following, safety, and clarity. • Detected hallucinations, completed preference annotation tasks, and performed multi-step Chain-of-Thought labeling.

2024 - Present

Show more

Education

Work History