Key Skills

Software

AWS SageMaker

Appen

Argilla

Clickworker

Datature

Labelbox

Prodigy

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Text

Top Task Types

Data Collection

Evaluation Rating

Text Summarization

Translation Localization

Freelancer Overview

I bring expertise in data analysis, labeling, and AI training data preparation, with a proven ability to organize, categorize, and annotate large datasets for predictive modeling and natural language processing. My background in marketing science and sentiment analysis allows me to capture subtle emotional tones, making labeled data not only technically accurate but also contextually meaningful. Skilled in data cleaning, categorization, and visualization, I ensure high-quality inputs that improve the reliability of AI models.

Labeling Experience

E-Commerce App Conversion Funnel

AppenTextRelationshipClassification

Labeled 75 user interaction sessions from a retail shopping app. Each session was annotated with funnel stages (browse, cart, checkout, purchase) and flagged for abandonment points. Data supported a small-scale statistical analysis of conversion rates, helping optimize A/B testing for in-app marketing.

Labeled 75 user interaction sessions from a retail shopping app. Each session was annotated with funnel stages (browse, cart, checkout, purchase) and flagged for abandonment points. Data supported a small-scale statistical analysis of conversion rates, helping optimize A/B testing for in-app marketing.

2025 - 2025

Mobile Fitness App Usage Logs

Labelbox3D SensorClassificationEvaluation Rating

Annotated 100 anonymized fitness app sessions to classify activity type (walking, running, cycling) and label anomalies in step count or calorie estimates. Data was validated with manual checks to ensure accuracy before being used in a pilot statistical model predicting daily activity levels.

Annotated 100 anonymized fitness app sessions to classify activity type (walking, running, cycling) and label anomalies in step count or calorie estimates. Data was validated with manual checks to ensure accuracy before being used in a pilot statistical model predicting daily activity levels.

2025 - 2025

Investment Commentary Sentiment Tagging

ProdigyTextClassificationEmotion Recognition

Labeled 80 short investment memos and commentaries from boutique asset managers. Each entry was tagged for sentiment (bullish, bearish, neutral), tone (cautious, confident, urgent), and key themes (market risk, opportunity, portfolio allocation). Summaries were created to condense analyst tone into structured labels. QA involved double-pass review to maintain consistent sentiment scoring, supporting the development of a model for predicting investor mood shifts.

Labeled 80 short investment memos and commentaries from boutique asset managers. Each entry was tagged for sentiment (bullish, bearish, neutral), tone (cautious, confident, urgent), and key themes (market risk, opportunity, portfolio allocation). Summaries were created to condense analyst tone into structured labels. QA involved double-pass review to maintain consistent sentiment scoring, supporting the development of a model for predicting investor mood shifts.

2025 - 2025

Bank Statement Risk Review

Scale AITextClassificationEmotion Recognition

Annotated 40 anonymized bank statements to flag irregular patterns for a financial risk analysis project. Tasks included labeling transaction type (recurring, one-time, transfer), identifying unusual cash flows, and rating transactions for potential fraud risk. Focused on high-precision annotation for boutique risk modeling, with results validated through spot audit checks for consistency.

Annotated 40 anonymized bank statements to flag irregular patterns for a financial risk analysis project. Tasks included labeling transaction type (recurring, one-time, transfer), identifying unusual cash flows, and rating transactions for potential fraud risk. Focused on high-precision annotation for boutique risk modeling, with results validated through spot audit checks for consistency.

2025 - 2025

Expense Receipt Classification

LabelboxTextEntity Ner ClassificationClassification

Annotated a boutique dataset of 95 scanned expense receipts for a small business accounting tool. Tasks included tagging merchant name, transaction date, amount, and expense category (travel, meals, supplies). Each receipt image was cross-checked against extracted text for consistency, with a final accuracy rate above 96%. Deliverables helped refine automated expense categorization for bookkeeping workflows.

Annotated a boutique dataset of 95 scanned expense receipts for a small business accounting tool. Tasks included tagging merchant name, transaction date, amount, and expense category (travel, meals, supplies). Each receipt image was cross-checked against extracted text for consistency, with a final accuracy rate above 96%. Deliverables helped refine automated expense categorization for bookkeeping workflows.

2025 - 2025

Show more

Education

Work History