Prosper O.

Key Skills

Software

Scale AI

Appen

Surge AI

CVAT

Labelbox

Don't disclose

Top Subject Matter

LLM fine-tuning and RLHF dataset annotation (NLP and multimodal labeling for AI clients)

Computer vision dataset labeling for autonomous driving and NLP companion datasets

NLP dataset annotation and quality assurance for intent/sentiment and benchmark creation

Top Data Types

Text

Image

Document

Top Task Types

RLHF

Object Detection

Evaluation/Rating

Data Collection

Prompt + Response Writing (SFT)

Question Answering

Text Summarization

Text Generation

Freelancer Overview

Senior Data Labeling Specialist & QA Lead (Remote AI Contractor). Core strengths include Scale AI, Appen, and Surge AI. Education includes Bachelor of Science, New York University (NYU) (2016) and Certificate, Coursera / DeepLearning.AI (2016). AI-training focus includes data types such as Text and Image and labeling workflows including RLHF, Object Detection, and Entity (NER).

Labeling Experience

Senior Data Labeling Specialist & QA Lead (Remote AI Contractor)

Scale AITextRLHF

Delivered 500K+ labeled data points monthly across NLP, vision, and multimodal tasks for Fortune 500 AI clients while meeting SLAs of 95%+. Specialized in RLHF annotation by ranking and rating LLM outputs for helpfulness, harmlessness, and honesty to support fine-tuning pipelines. Authored guidelines and edge-case decision trees and performed IAA audits to ensure label consistency and reduce ambiguity. • Ranked and rated LLM responses for RLHF categories (helpfulness, harmlessness, honesty) • Built and maintained annotation guidelines and edge-case decision trees for 10+ project types • Conducted IAA audits using Cohen's Kappa, maintaining scores above 0.90 • Executed systematic QA sampling at 10–20% to catch and correct label errors before delivery

2022 - Present

Data Annotation Team Lead

CVATImageObject Detection

Led a 12-person annotation team producing high-volume computer vision and NLP datasets with weekly output targets of 50K+ labels. Implemented a 5-tier QA scoring framework to reduce client rejection rates by 45% and trained annotators to improve labeling consistency. Managed annotation delivery for a 200K-image autonomous driving dataset using bounding boxes, lane markings, and pedestrian keypoints. • Managed daily annotation output targets for multi-project computer vision and NLP labeling • Trained and onboarded annotators on CVAT, Labelbox, and Appen workflows • Built SQL-based audit pipelines to detect class imbalance, annotation drift, and outliers • Delivered autonomous driving image annotations with bounding boxes and keypoints at a 96% quality pass rate

2019 - 2021

Junior Data Annotator & Analyst

Don't discloseText

Annotated 300K+ text and image samples for NLP intent classification, sentiment analysis, and object detection projects. Used SQL to query and validate datasets, identifying mislabeled samples and inconsistencies in training batches. Supported evaluation by creating gold-standard test sets and benchmark labels for QA benchmarking pipelines. • Performed text annotation for intent classification and sentiment analysis • Validated and cleaned training data using SQL-based checks • Helped develop internal annotation standards and quality checklists • Created gold-standard benchmark labels for model evaluation and QA pipelines

2016 - 2019

Education

Work History