Shafinaz Keply

Improved the accuracy of speech recognition systems through detailed transcription and data labeling tasks. I listened to short English audio clips, categorized them by speaker, identified accents, and accurately transcribed spoken words into clear written text. Every clip required careful attention to tone, pronunciation, and context to ensure that the final output reflected natural and human-like language.

I also reviewed my work for consistency and corrected any transcription errors before submission. The project demanded focus, precision, and the ability to handle a high volume of data while maintaining quality. Working remotely under a flexible, task-based system helped me develop strong time management and organization skills.

This experience deepened my understanding of how clean and accurate data contributes to building more innovative AI tools and speech recognition technologies.

Enhanced the authenticity and fluency of AI-generated Malay text through detailed post-editing. I reviewed and refined existing text units so that they sounded natural, accurate, and contextually appropriate for native Malay speakers. Each edit required careful attention to tone, flow, and meaning while preserving the intended message of the original output.

I worked with text data delivered in spreadsheet format, making systematic edits to ensure linguistic quality, readability, and cultural relevance. The goal was to enhance the naturalness of the text while maintaining consistency across multiple content types.

Through this project, I strengthened my ability to assess AI-generated content critically, apply linguistic judgment, and enhance model output quality through precise, human-verified editing.

This project involved labeling image datasets to support the development and training of computer vision models. Using the CVAT tool, I created bounding box annotations around target objects to help AI systems recognize and classify visual elements more accurately. My primary focus was on ensuring precise and consistent labeling in accordance with project instructions.

Throughout the process, I practiced applying annotation guidelines, naming conventions, and quality checks to maintain accuracy. I also learned how structured annotation data contributes to model performance and why attention to detail is essential in image-based AI projects.

This experience helped me strengthen my understanding of annotation workflows, data organization, and the importance of precision in real-world AI training tasks.