AI Tester, Search Query and Ad Safety Evaluator, Data Annotator
Hello! I’m Arnab Saba, an AI evaluator and data annotation specialist with hands-on experience in testing cutting-edge AI technologies.
I have spent over two year working with AI models, refining responses, evaluating performance, and ensuring AI-generated content meets quality standards. My work spans multiple platforms, including Gemini, Outlier, Toloka, Appen, and Meta's Avatar Testing Community. I thrive on providing insightful feedback that helps improve AI systems and online platforms.
Beyond AI testing, I have also contributed to data annotation, ad relevance evaluation, and text-to-speech assessments, ensuring accuracy and relevance across various projects. I have participated in several groundbreaking AI-based programs, gaining firsthand experience in evaluating emerging technologies.
Key Experience & Projects
- AI Testing & Training (Gemini & Outlier)
- Over a year of experience in evaluating AI-generated responses and improving model performance.
- Currently working on AI training in Urdu, focusing on prompt writing, response analysis, and feedback refinement.
- AI Evaluation & Data Annotation
- Toloka: Expertise in data labeling, text-to-speech evaluation, and search relevance assessment.
- Appen: Ad relevance & safety evaluation, text classification, and content moderation.
- Meta Avatar Testing: Provided detailed usability and accuracy feedback on digital avatars.
- Voice Recording & Linguistic Tasks
- Completed voice recording gigs for platforms like Voices & Appen, with all tasks approved.
- Website & UX Testing
- Participated in several survey-based projects focused on evaluating small interface elements, website usability, and user experience feedback.
Skills
- AI Testing & Evaluation
- Data Annotation & Labeling
- AI Prompt Writing & Response Review
- Ad Relevance & Safety Evaluation
- Text-to-Speech Evaluation
- Multilingual: English, Hindi, and Urdu
Education
- Bachelor’s Degree in Political Science
- Currently Pursuing a Master’s in Political Science
I’m always excited to collaborate on projects that push AI and technology forward. If you're looking for reliable, detail-oriented insights to enhance AI models or digital platforms, let’s connect!
Work Terms
I am available to work 30 hours a week, but i can adjust it on the basis of client's and work demand, and i am open to discuss on any aspect of the project.