Miraniaina Ramananjanahary

Key Skills

Software

Label Studio

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Entity Ner Classification

Freelancer Overview

I am an operational support professional with hands-on experience in digital marketing, lead generation, and AI-assisted communication. My background includes executing targeted outreach campaigns using AI prompts, managing large datasets for prospecting, and utilizing tools like ChatGPT, AI Studio, and Gemini to optimize engagement and workflow. I am skilled at working independently in remote environments, leveraging platforms such as Notion, Google Workspace, and Microsoft Teams to ensure seamless collaboration and data management. My ability to draft effective outreach messages and manage CRM data highlights my attention to detail and adaptability—key qualities for data labeling and AI training data roles. I am eager to contribute my expertise in data annotation, process optimization, and technology-driven projects to support high-quality AI development.

Labeling Experience

Intent Classification and Entity Extraction for E-commerce Customer Support

Label StudioTextEntity Ner Classification

The project aimed to train a Large Language Model (LLM) to automate the sorting of customer requests. My role involved analyzing 500+ messages to: Intent Classification: Categorizing whether a message was a refund request, shipping inquiry, or technical complaint. Entity Extraction: Precisely highlighting order numbers (e.g., #12345), product names, and dates mentioned in the text. Sentiment Analysis: Assigning a sentiment score from -1 (angry/dissatisfied) to +1 (satisfied) to help the AI prioritize urgent escalations.

The project aimed to train a Large Language Model (LLM) to automate the sorting of customer requests. My role involved analyzing 500+ messages to: Intent Classification: Categorizing whether a message was a refund request, shipping inquiry, or technical complaint. Entity Extraction: Precisely highlighting order numbers (e.g., #12345), product names, and dates mentioned in the text. Sentiment Analysis: Assigning a sentiment score from -1 (angry/dissatisfied) to +1 (satisfied) to help the AI prioritize urgent escalations.

2025 - 2025

Education

Work History