Peter W.

Key Skills

Software

iMerit

Appen

Other

Top Subject Matter

Image quality assessment for AI training datasets

Image pair labeling and rubric-based validation for AI training

Multilingual legal document AI (RAG) annotation and evaluation

Top Data Types

Image

Document

Text

Top Task Types

Entity (NER) Classification

Data Collection

Classification

Mapping

Transcription

Freelancer Overview

Remote Image Quality Rater – Nidra WF11. Brings 10+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include iMerit, Appen, and Internal. Education includes Bachelor of Business Information Technology, Mount Kenya University (2024) and Diploma in Artificial Intelligence, Advanced Solutions Technical Institute (AICE Africa) (2022). AI-training focus includes data types such as Image, Document, and Geospatial and labeling workflows including Evaluation, Rating, and Entity (NER) Classification.

Labeling Experience

AI Research Engineer (Volunteer) – Legal AI Africa

DocumentEntity (NER) Classification

Contribute to a production multilingual RAG-based legal assistant by developing structured annotation guidelines and model evaluation frameworks for low-resource African language contexts. Support the ingestion of legal document chunks into a pgvector knowledge base and contribute to responsible AI-aligned model evaluation. Apply structured annotation approaches to improve retrieval and response quality for francophone African users. • Develop annotation guidelines for multilingual legal data • Support model evaluation framework contributions • Assist with documentation and structured evaluation for RAG • Contribute to multilingual legal assistant readiness

2026 - Present

Remote Image Quality Rater – Nidra WF11

iMeritImage

Rate image quality using predefined IceCow evaluation rubrics for structured criteria including sharpness, blur, composition, and lighting. Ensure guideline adherence across high-volume batches and flag ambiguous or inconsistent cases to protect downstream AI training reliability. Maintain consistent, independent quality within platform accuracy thresholds while completing repetitive review tasks. • Apply sharpness/blur/composition/lighting scoring per rubric • Flag low-quality or ambiguous images for dataset integrity • Perform independent reviews within structured expectations • Deliver consistent quality within multiple project batches

2025 - Present

ML Engineering Support (UNV Online Volunteer)

OtherData Collection

Aggregate and preprocess multi-source environmental datasets including satellite imagery and sensor networks for climate risk modeling workflows. Perform exploratory data analysis using structured analytical rubrics to assess distributions, outliers, and feature relevance for model readiness. Conduct research and synthesis on AI foundation models (GenCast, GraphCast, Prithvi) and produce structured documentation mapping key variables and data requirements. • Build Python-based preprocessing pipelines for multi-source datasets • Conduct rubric-driven EDA for model-readiness assessment • Research foundation models and document variable/data requirements • Deliver ML-ready, high-integrity datasets for forecasting initiatives

2025 - 2025

Data Labeling Specialist – Project Coffee

AppenImage

Validate, tag, and complete image-pair labeling tasks for Project Coffee, a high-volume fast-turnaround AI data labeling initiative. Perform image editing conversation tasks and evaluate product image pairs using task-specific rubrics to ensure consistent outputs. Qualify through calibration and assessment (English comprehension and oral assessment) before receiving full project access and maintaining annotation standards. • Complete validation and tagging for high-throughput labeling • Perform image-pair evaluations with task rubrics • Complete conversation-style image editing tasks • Meet Appen/Crowdgen quality benchmarks under speed and accuracy constraints

2025 - 2025

Case Processing Assistant

DocumentClassification

Process refugee resettlement cases by applying strict inter-agency data standards and verification protocols within a START Case Management System. Ensure accurate, structured recording of case information as high-stakes annotated data used for downstream case workflows. Produce analysis and reporting in Tableau and Excel to surface case trends, bottlenecks, and workflow inefficiencies while maintaining sensitive information security compliance. • Apply verification protocols for 500+ USRAP cases • Maintain structured case data standards in START system • Conduct trend and bottleneck reporting in Tableau/Excel • Maintain data security compliance for sensitive refugee information

2022 - 2023

Show more

Education

Work History