Yahya A.

Key Skills

Software

Scale AI

Other

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Video

Top Task Types

Computer Programming/Coding

Evaluation/Rating

Prompt + Response Writing (SFT)

RLHF

Freelancer Overview

An experienced specialist in video and image data annotation, having contributed to a diverse range of projects through platforms such as Amazon Mechanical Turk. My work has supported the development of AI and machine learning models by ensuring high-quality, accurately labeled datasets. In addition to my annotation expertise, I am currently focused on computer programming and front-end development for both mobile and web applications. I have hands-on experience with modern technologies including Swift, Python, React, and TypeScript, and have collaborated with leading organizations such as Scale AI (Outlier) to further enhance data quality and AI solutions.

Labeling Experience

Andromeda UI - Providing Golden Response for SOTA Model

Scale AIComputer Code ProgrammingRLHFEvaluation/Rating

This project's goal was to create websites based on seeded prompts for state-of-the-art (SOTA) models, ensuring that outputs meet Fortune 500 company standards. We were asked to rewrite the seeded prompts to be more natural, reflecting how humans might query the model. We then ranked the model responses based on three criteria: visual design, functionality, and instruction following, providing detailed explanations within the evaluation rubric. The core of this project involved providing a golden response based on the selected SOTA model response. This ensured that the final output would match or closely approximate Fortune 500 company standards, with no UI bugs, excellent functionality, and best-practice coding standards. I utilized React, TypeScript, Next.js, Tailwind CSS, HTML, and CSS for this project.

2025

Hopper - Providing Indonesian Localization for SOTA Models

Scale AIComputer Code ProgrammingTranslation/LocalizationRLHF

Rewrite the seeded response from English to my native language (Indonesian) and make it more natural, in a way that a real human might ask. Then, ensure the response returns correct localization where applicable, including dates, currencies, variable naming, and code comments. After that, provide ratings to the model based on the localization rubric.

2025 - 2025

Expert Sturgeon - Model Stomping Using Swift

Scale AIComputer Code ProgrammingRLHFEvaluation/Rating

In this project, we were tasked with creating prompts specifically designed to challenge AI models, ensuring that at least one of the four models would fail to meet the standards in one of the given rubric categories. The evaluation criteria varied depending on the specific categories, but generally required us to assess correctness, completeness, clarity, readability, performance, best practices, and visual design. I primarily used Swift for this project.

2025 - 2025

Video Categorization

OtherVideoClassificationData Collection

Give tagging and classification to the given video.

2016 - 2019

Education

Work History