For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
Christina F

Christina F

Cybersecurity Analyst - Healthcare IT and AI

Land O Lakes, A
Entry LevelInternal Proprietary ToolingDon T Disclose

Key Skills

Software

Internal/Proprietary Tooling
Don't disclose

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code ProgrammingComputer Code Programming
TextText

Top Label Types

Bounding Box
Classification
Computer Programming Coding
Data Collection
Evaluation Rating
Prompt Response Writing SFT
Question Answering
Red Teaming
Text Generation
Text Summarization
Transcription
Translation Localization

Freelancer Overview

I bring hands-on experience in AI training data and data annotation, having worked as an AI Trainer for Handshake AI, where I developed and evaluated domain-specific prompts to assess large language model performance in cybersecurity. My background includes analyzing LLM outputs for accuracy and clarity, providing expert feedback to improve AI understanding of complex topics, and conducting independent research to support high-quality prompt development. I am skilled in documentation, technical analysis, and cross-functional collaboration, with a strong foundation in cybersecurity tools and concepts. My experience spans both technical and client-facing roles, ensuring clear communication and attention to detail in all aspects of data labeling and annotation.

Entry LevelEnglish

Labeling Experience

Bounding Box Annotation — People & Objects

Internal Proprietary ToolingImageBounding BoxClassification
I completed bounding box annotation tasks for both person-level and general object localization. For person annotation, I sourced large group photographs containing at least ten clearly visible individuals, drew tight bounding boxes around each person with attention to face and upper body coverage, assigned unique contextual background settings for AI image generation, and evaluated the quality of generated portraits against the original subjects — checking for facial feature accuracy, outfit consistency, natural pose variation, and absence of artifacts. For object annotation, I identified target objects from prompt descriptions, drew precise boxes with minimal excess space, and verified edge accuracy using zoom-based quality checks.

I completed bounding box annotation tasks for both person-level and general object localization. For person annotation, I sourced large group photographs containing at least ten clearly visible individuals, drew tight bounding boxes around each person with attention to face and upper body coverage, assigned unique contextual background settings for AI image generation, and evaluated the quality of generated portraits against the original subjects — checking for facial feature accuracy, outfit consistency, natural pose variation, and absence of artifacts. For object annotation, I identified target objects from prompt descriptions, drew precise boxes with minimal excess space, and verified edge accuracy using zoom-based quality checks.

2025

Event Timestamps & Scene Cut Detection

Internal Proprietary ToolingVideoPoint Key PointSegmentation
I reviewed AI-detected scene cuts in video content and verified, corrected, and supplemented them with manually identified segment boundaries. Each cut marker had to be placed on the exact first frame of a new scene or event, requiring frame-by-frame analysis using timeline scrubbing tools. Beyond verifying camera cuts, I further segmented footage based on logical event boundaries — distinct actions or moments occurring within the same camera angle — and applied concise, descriptive labels to each segment. This task demanded sustained focus, precise temporal judgment, and the ability to distinguish meaningful content transitions from irrelevant movement.

I reviewed AI-detected scene cuts in video content and verified, corrected, and supplemented them with manually identified segment boundaries. Each cut marker had to be placed on the exact first frame of a new scene or event, requiring frame-by-frame analysis using timeline scrubbing tools. Beyond verifying camera cuts, I further segmented footage based on logical event boundaries — distinct actions or moments occurring within the same camera angle — and applied concise, descriptive labels to each segment. This task demanded sustained focus, precise temporal judgment, and the ability to distinguish meaningful content transitions from irrelevant movement.

2025

Multi-Dimensional AI Output Comparison (H2H Evaluation)

Internal Proprietary ToolingVideoClassificationQuestion Answering
I conducted side-by-side comparisons of AI-generated image and video outputs, evaluating each pair across multiple independent dimensions including instruction following, visual quality, video-audio alignment, and AI artifact detection. I assessed outputs for issues such as lip sync accuracy, anatomical inconsistencies, text rendering errors, and physical impossibilities before forming an overall holistic judgment. This structured, dimension-first approach ensured that my evaluations were consistent and evidence-based rather than driven by first impressions, resulting in high-quality feedback that contributed to improving model performance.

I conducted side-by-side comparisons of AI-generated image and video outputs, evaluating each pair across multiple independent dimensions including instruction following, visual quality, video-audio alignment, and AI artifact detection. I assessed outputs for issues such as lip sync accuracy, anatomical inconsistencies, text rendering errors, and physical impossibilities before forming an overall holistic judgment. This structured, dimension-first approach ensured that my evaluations were consistent and evidence-based rather than driven by first impressions, resulting in high-quality feedback that contributed to improving model performance.

2025

Diagram Label Removal Audit

Internal Proprietary ToolingImageBounding BoxEntity Ner Classification
I conducted detailed audits of AI-generated diagram label removal tasks. For each submission, I evaluated whether all visible text labels had been correctly removed while preserving key metadata such as titles, subtitles, and logos. I applied a structured four-point rating scale — ranging from Major Issues to Exceptional — to assess output quality, and provided specific written feedback to guide model improvement. This work required strong visual attention to detail and the ability to distinguish between intentional content removal and AI errors such as residual artifacts or incorrectly altered elements.

I conducted detailed audits of AI-generated diagram label removal tasks. For each submission, I evaluated whether all visible text labels had been correctly removed while preserving key metadata such as titles, subtitles, and logos. I applied a structured four-point rating scale — ranging from Major Issues to Exceptional — to assess output quality, and provided specific written feedback to guide model improvement. This work required strong visual attention to detail and the ability to distinguish between intentional content removal and AI errors such as residual artifacts or incorrectly altered elements.

2025

Ultra-Dense & Essence-Based Image Captioning

Don T DiscloseImageBounding BoxSegmentation
I completed both ultra-dense and essence-based image captioning tasks. For ultra-dense captions, I produced exhaustive written descriptions covering every visible element of an image — including objects, spatial relationships, lighting, textures, camera angle, and image quality — written in sufficient detail that someone could understand or recreate the image without seeing it. For essence-based captions, I distilled images to their most important elements in concise, memorable language. Both task types required independent, original writing without AI assistance, testing my descriptive accuracy, vocabulary range, and ability to shift between comprehensive and summary-level communication.

I completed both ultra-dense and essence-based image captioning tasks. For ultra-dense captions, I produced exhaustive written descriptions covering every visible element of an image — including objects, spatial relationships, lighting, textures, camera angle, and image quality — written in sufficient detail that someone could understand or recreate the image without seeing it. For essence-based captions, I distilled images to their most important elements in concise, memorable language. Both task types required independent, original writing without AI assistance, testing my descriptive accuracy, vocabulary range, and ability to shift between comprehensive and summary-level communication.

2025

Education

W

Western Governors University

Master of Science, Cybersecurity

Master of Science
2024 - 2025
S

SANS Women's Immersion Academy

Certificate (GIAC Certification Program), Cybersecurity

Certificate (GIAC Certification Program)
2021 - 2022

Work History

H

Handshake AI Fellowship

Cybersecurity Domain Expert - AI Trainer

Land O' Lakes
2025 - Present
S

St. Andrew Post-Acute Rehabilitation

Licensed Physical Therapist Assistant

Tampa, FL
2023 - Present