Emma Onsongo

Key Skills

Software

Labelbox

Surge AI

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Image

Text

Video

Top Task Types

Bounding Box

Classification

Data Collection

Entity Ner Classification

Point Key Point

Prompt Response Writing SFT

Segmentation

Text Generation

Text Summarization

Tracking

Freelancer Overview

A results-driven professional with over 10 years of human rights advocacy and research experience, 5 years in legal practice, 3+ years experience in data labeling and AI quality assurance, and clinical behavioral support experience in healthcare. Experienced in applying sophisticated legal and policy analysis, litigation and client advocacy, and high-accuracy annotation of large-scale text and multimodal datasets to machine learning systems. Proven ability in performing quality audits, calibrating processes, and reviewing sensitive content, consistently meeting strict performance and accuracy benchmarks. Recognized for proficiency in analytical thinking, ethical judgment, strong decision-making, leadership, and the ability to work effectively both independently and within cross-functional teams. Passionate about leveraging research, technology, and community-centered practice to drive organizational growth, data integrity, and meaningful social impact.

Labeling Experience

Banking Text Data Annotation & Quality Review Project

Surge AITextBounding BoxPoint Key Point

Built on prior knowledge of large-scale text annotation and quality assurance projects to assist in training large language models. Conducted annotation in Prodigy of complex prompts and user queries, as well as AI-generated responses for intent, sentiment, named entities, and response quality in accordance with detailed annotation guidelines. Analyzed ambiguous cases and edge scenarios, documented the reasoning behind them, and provided examples to promote guideline clarity. Performed secondary reviews of teammates' work, identified recurring error patterns, and provided clear feedback to maintain team standards. Engaged in calibration sessions with project managers and QA teams for consistent labeling standards. Assisted in sensitive content classification activities by carefully applying safety policies and escalation procedures.

Built on prior knowledge of large-scale text annotation and quality assurance projects to assist in training large language models. Conducted annotation in Prodigy of complex prompts and user queries, as well as AI-generated responses for intent, sentiment, named entities, and response quality in accordance with detailed annotation guidelines. Analyzed ambiguous cases and edge scenarios, documented the reasoning behind them, and provided examples to promote guideline clarity. Performed secondary reviews of teammates' work, identified recurring error patterns, and provided clear feedback to maintain team standards. Engaged in calibration sessions with project managers and QA teams for consistent labeling standards. Assisted in sensitive content classification activities by carefully applying safety policies and escalation procedures.

2023 - 2025

Text Data Curation

LabelboxTextText GenerationTracking

Parsed and hand-curated large collections of text data to prepare them for training customer interaction tasks. Conducted quality labeling assurance with consistency check and multi-level checking to improve dataset integrity and minimize errors.

Parsed and hand-curated large collections of text data to prepare them for training customer interaction tasks. Conducted quality labeling assurance with consistency check and multi-level checking to improve dataset integrity and minimize errors.

2023 - 2023

Image Annotation for Computer Vision Model Training

AppenImageBounding BoxPolygon

The use of annotated large-scale image datasets to train a computer vision model was done by drawing bounding boxes and polygons of vehicles, pedestrians, storefronts, products, and signage. Conducted semantic segmentation for mapping roads, sidewalks, background objects, and other scene elements. Maintained strict annotation guidelines and quality rubrics as per accuracy benchmarks. Attended calibration sessions and fixed low-quality annotations based on QA feedback. This work helped develop object detection and scene understanding models.

The use of annotated large-scale image datasets to train a computer vision model was done by drawing bounding boxes and polygons of vehicles, pedestrians, storefronts, products, and signage. Conducted semantic segmentation for mapping roads, sidewalks, background objects, and other scene elements. Maintained strict annotation guidelines and quality rubrics as per accuracy benchmarks. Attended calibration sessions and fixed low-quality annotations based on QA feedback. This work helped develop object detection and scene understanding models.

2021 - 2022

Education

Work History