Francis Karanja

Key Skills

Software

Prodigy

LabelImg

Top Subject Matter

No subject matter listed

Top Data Types

Document

Text

Top Label Types

Classification

Freelancer Overview

I am an experienced data analyst and OSINT researcher specializing in high-integrity data verification, annotation, and manual auditing. My background includes orchestrating large-scale manual audits of over 2,000 global legal entities, where I meticulously labeled, validated, and reconciled complex corporate data across multiple jurisdictions. I excel at synthesizing fragmented datasets, identifying discrepancies, and ensuring dataset integrity through advanced logical reasoning and human-in-the-loop verification processes. I am highly proficient with data management tools such as MS Excel, Google Sheets, and specialized OSINT platforms like Open corporates and Crunchbase. My hands-on approach to data quality—especially in environments where automated tools fall short equips me to deliver accurate, reliable datasets for AI training, legal entity resolution, and analytical projects.

Labeling Experience

Lead OSINT Researcher & Data Auditor

ProdigyDocumentClassification

Oversaw exhaustive manual verification and classification of global legal entity datasets for corporate intelligence purposes. Used human-in-the-loop techniques to label and validate director and shareholder information extracted from registry documents. Labeled data by distinguishing between confirmed, unconfirmed, and potentially fraudulent or erroneous entries using advanced audit trail methods. • Conducted manual entity validation across international registry datasets. • Utilized pattern recognition and logical reasoning for document classification. • Applied audit trail techniques for label reliability and traceability. • Focused on data accuracy, mitigating risk of erroneous or hallucinated outputs from automation.

Oversaw exhaustive manual verification and classification of global legal entity datasets for corporate intelligence purposes. Used human-in-the-loop techniques to label and validate director and shareholder information extracted from registry documents. Labeled data by distinguishing between confirmed, unconfirmed, and potentially fraudulent or erroneous entries using advanced audit trail methods. • Conducted manual entity validation across international registry datasets. • Utilized pattern recognition and logical reasoning for document classification. • Applied audit trail techniques for label reliability and traceability. • Focused on data accuracy, mitigating risk of erroneous or hallucinated outputs from automation.

2023

Junior Data Analyst

LabelimgTextClassification

Transformed unstructured qualitative survey responses into classified quantitative data sets for analysis. Labeled and validated respondent data to ensure statistical reliability of outputs. Used human review to implement multi-stage verification and logical outlier detection in labeling workflows. • Converted textual survey responses to classified categories for data analysis. • Implemented human QA on data labeling to improve precision. • Detected labeling anomalies and logical outliers using multi-stage checks. • Collaborated with QA teams to refine annotation/labeling processes.

Transformed unstructured qualitative survey responses into classified quantitative data sets for analysis. Labeled and validated respondent data to ensure statistical reliability of outputs. Used human review to implement multi-stage verification and logical outlier detection in labeling workflows. • Converted textual survey responses to classified categories for data analysis. • Implemented human QA on data labeling to improve precision. • Detected labeling anomalies and logical outliers using multi-stage checks. • Collaborated with QA teams to refine annotation/labeling processes.

2021 - 2022

Education

Work History