Monica G.

Key Skills

Software

Scale AI

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Task Types

Audio Recording

Emotion Recognition

Evaluation/Rating

Object Detection

Prompt + Response Writing (SFT)

RLHF

Transcription

Freelancer Overview

I am an AI Trainer and Data Specialist with hands-on experience in data labeling, annotation, and evaluation for both computer vision and NLP projects. My work includes structured assessment of computer vision face swap outputs, in-depth analysis of NLP audio transcriptions, and preference ranking for AI-generated responses across 30+ projects. I have a strong background in prompt engineering, synthetic prompt testing, and data quality assurance, consistently optimizing AI model outputs for accuracy, consistency, and reliability. With a foundation in psychology and human resources, I bring an analytical approach and a passion for user-centric, ethical AI solutions to every project I take on.

Labeling Experience

AI Data Specialist (Computer Vision & NLP)

Internal Proprietary ToolingImageEmotion RecognitionObject Detection

Ruby Project – Computer Vision (Faceswap Capability) • Evaluated 190+ model outputs by reviewing source images, face media, and final renders to determine Pass/Not Pass status. • Identified visual anomalies, including mismatched identities, digital artifacts, expression inconsistencies, and subtle signs of image manipulation. • Classified rejection reasons into predefined categories to provide structured feedback for model optimization. Echo Project – Natural Language Processing (Audio Transcript Review) • Conducted quality assessment for the Echo Project, specializing in Audio Transcript Review and NLP accuracy. • Evaluated and audited 360+ audio samples against generated transcripts to ensure linguistic precision, context alignment, and transcription integrity. • Analyzed audio-to-text outputs to identify grammatical errors, misheard words, and formatting inconsistencies.

Ruby Project – Computer Vision (Faceswap Capability) • Evaluated 190+ model outputs by reviewing source images, face media, and final renders to determine Pass/Not Pass status. • Identified visual anomalies, including mismatched identities, digital artifacts, expression inconsistencies, and subtle signs of image manipulation. • Classified rejection reasons into predefined categories to provide structured feedback for model optimization. Echo Project – Natural Language Processing (Audio Transcript Review) • Conducted quality assessment for the Echo Project, specializing in Audio Transcript Review and NLP accuracy. • Evaluated and audited 360+ audio samples against generated transcripts to ensure linguistic precision, context alignment, and transcription integrity. • Analyzed audio-to-text outputs to identify grammatical errors, misheard words, and formatting inconsistencies.

2025

AI Trainer

Scale AITextRLHF

• Evaluated and refined AI-generated responses across 32+ diverse projects, focusing on enhancing model logic, factual reliability, and instruction-following capabilities. • Conducted rigorous factual accuracy audits by verifying model outputs against authoritative data sources to identify and mitigate "hallucinations" or misinformation. • Performed synthetic prompt evaluation and preference ranking, analyzing multiple model outputs to select the highest-quality response based on accuracy, safety, and helpfulness. • Developed and tested prompt engineering techniques, including the use of detailed constraints and personas, to observe how different instruction structures impact the precision of the model’s performance. • Contributed to the Reinforcement Learning from Human Feedback (RLHF) process, providing structured data to help the model better understand complex human nuances and intent.

• Evaluated and refined AI-generated responses across 32+ diverse projects, focusing on enhancing model logic, factual reliability, and instruction-following capabilities. • Conducted rigorous factual accuracy audits by verifying model outputs against authoritative data sources to identify and mitigate "hallucinations" or misinformation. • Performed synthetic prompt evaluation and preference ranking, analyzing multiple model outputs to select the highest-quality response based on accuracy, safety, and helpfulness. • Developed and tested prompt engineering techniques, including the use of detailed constraints and personas, to observe how different instruction structures impact the precision of the model’s performance. • Contributed to the Reinforcement Learning from Human Feedback (RLHF) process, providing structured data to help the model better understand complex human nuances and intent.

2024 - 2025

Education

Work History