Kareema G.

Key Skills

Software

Telus

Other

Top Subject Matter

Multimodal video action labeling and model output evaluation (video+audio).

LLM output scoring and voice/speech quality evaluation.

AI education content development (prompting guidance).

Top Data Types

Video

Audio

Text

Document

Top Task Types

Action Recognition

Prompt + Response Writing (SFT)

Freelancer Overview

Video Annotation & AI Training Contractor, Telus International (Remote). Brings 5+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Telus and Other. Education includes Master of Laws, University of South Florida (2025) and Bachelor of Laws, Florida State University (2024). AI-training focus includes data types such as Video, Audio, and Text and labeling workflows including Action Recognition, Evaluation, and Rating.

Labeling Experience

Video Annotation & AI Training Contractor, Telus International (Remote)

TelusVideoAction Recognition

Performed video action labeling, bounding boxes, frame segmentation, timestamping, and scene classification to support multimodal dataset creation and QA. Conducted guideline-aligned QA checks to maintain annotation consistency across samples. Evaluated video+audio model outputs for clarity, accuracy, and visual context, and escalated edge cases to improve taxonomy and model behavior. • Labeled 500+ high-precision video instances with >95% consistency scores on platform QA metrics. • Supported model performance improvements by escalating difficult cases and contributing feedback to taxonomy updates. • Applied structured guideline compliance during annotation and post-annotation QA. • Reviewed multimodal outputs (visual/audio) to assess end-to-end usefulness for downstream tasks.

2025 - Present

LLM / Voice Model Evaluator, Outlier i.a

OtherAudio

Evaluated speech synthesis and voice model outputs for prosody, tone, clarity, and human-likeness. Scored LLM outputs, completed prompt-following benchmarks, and performed pairwise/ranked response evaluation. Identified and flagged hallucinations, factual errors, toxic or unsafe output, and instruction violations. • Completed 2,000+ LLM and speech evaluation tasks using structured feedback reports. • Ranked responses and scored outputs to quantify quality and alignment issues. • Flagged safety and compliance risks including toxicity, unsafe content, and instruction non-adherence. • Provided evaluation findings to increase model accuracy over iterative improvement cycles.

2024 - 2025

AI Content Creator & Partner, Flowmingo.ai

OtherTextPrompt + Response Writing (SFT)

Produced AI-usage and tech communication content focused on prompting and practical guidance for end users. Developed educational materials that support user onboarding and improved engagement for AI-related product audiences. • Authored instructional content on how to prompt and use AI effectively. • Created engagement-focused educational posts for community onboarding. • Communicated AI concepts clearly for non-technical audiences. • Supported growth initiatives through content delivery aligned with product adoption goals.

2023 - 2024

Education

Work History