Princess Ikechi

Key Skills

Software

Labelbox

Scale AI

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Text

Top Label Types

Question Answering

RLHF

Evaluation Rating

Red Teaming

Prompt Response Writing SFT

Freelancer Overview

I have experience working on AI training data and data labeling projects that require high accuracy, consistency, and contextual understanding. My work has included annotating text for intent classification, sentiment analysis, entity recognition, and content moderation, as well as evaluating AI-generated responses for relevance, safety, and personalization quality. I am comfortable following detailed annotation guidelines, applying taxonomy rules, and maintaining inter-annotator consistency while meeting strict quality benchmarks and turnaround times. What sets me apart is my strong analytical thinking, attention to detail, and ability to balance precision with scalability. I have worked on projects involving conversational AI evaluation, data categorization, and quality assurance, where I identified edge cases, flagged ambiguous prompts, and provided structured feedback to improve model performance. With a background in structured documentation, executive support, and digital systems, I bring organization, critical reasoning, and clear communication to AI training workflows.

Labeling Experience

Audio Transcription & Speech Data Annotation for ASR Model Training

AppenAudioEmotion RecognitionEvaluation Rating

Contributed to large-scale audio annotation projects for Automatic Speech Recognition (ASR) and conversational AI. Tasks included verbatim and clean transcription of multi-speaker conversations, call center recordings, and voice assistant prompts, along with speaker diarization, timestamp alignment, and background noise labeling. Performed emotion recognition tagging, intent identification, and quality rating of model-generated transcriptions. Followed strict linguistic guidelines for punctuation, disfluencies, accent variations, and code-switching. Processed thousands of audio clips across diverse accents and environments, maintaining 95%+ accuracy. Ensured quality through double-pass reviews, calibration, spot audits, and strict SLA adherence.

Contributed to large-scale audio annotation projects for Automatic Speech Recognition (ASR) and conversational AI. Tasks included verbatim and clean transcription of multi-speaker conversations, call center recordings, and voice assistant prompts, along with speaker diarization, timestamp alignment, and background noise labeling. Performed emotion recognition tagging, intent identification, and quality rating of model-generated transcriptions. Followed strict linguistic guidelines for punctuation, disfluencies, accent variations, and code-switching. Processed thousands of audio clips across diverse accents and environments, maintaining 95%+ accuracy. Ensured quality through double-pass reviews, calibration, spot audits, and strict SLA adherence.

2024

Conversational AI Evaluation & RLHF Annotation for Large Language Models

Scale AITextQuestion AnsweringRLHF

Worked on large-scale conversational AI training and evaluation projects focused on improving response quality, safety, and alignment in LLM systems. My responsibilities included ranking model outputs, rating responses based on helpfulness, truthfulness, coherence, tone, and policy compliance, and identifying hallucinations or unsafe content. I applied RLHF (Reinforcement Learning from Human Feedback) guidelines to compare multiple AI responses and select the most aligned output. I also performed red teaming tasks to test model robustness by probing edge cases, ambiguous prompts, and adversarial inputs. The project involved annotating thousands of multi-turn conversations while maintaining high inter-annotator agreement and strict quality benchmarks (95%+ QA accuracy targets). Quality control measures included secondary reviews, calibration sessions, detailed rubric adherence, and structured feedback reporting to improve annotation consistency and reduce bias.

Worked on large-scale conversational AI training and evaluation projects focused on improving response quality, safety, and alignment in LLM systems. My responsibilities included ranking model outputs, rating responses based on helpfulness, truthfulness, coherence, tone, and policy compliance, and identifying hallucinations or unsafe content. I applied RLHF (Reinforcement Learning from Human Feedback) guidelines to compare multiple AI responses and select the most aligned output. I also performed red teaming tasks to test model robustness by probing edge cases, ambiguous prompts, and adversarial inputs. The project involved annotating thousands of multi-turn conversations while maintaining high inter-annotator agreement and strict quality benchmarks (95%+ QA accuracy targets). Quality control measures included secondary reviews, calibration sessions, detailed rubric adherence, and structured feedback reporting to improve annotation consistency and reduce bias.

2024

Education

Work History