Nurdyne Bank

Key Skills

Software

Scale AI

Top Subject Matter

Large Language Model (LLM) evaluation

Rlhf Domain Expertise

LLM creative text evaluation

Top Data Types

Text

Image

Document

Top Task Types

RLHF

Classification

Freelancer Overview

Freelance AI Data Annotator at Outlier (Scale AI), specializing in RLHF preference ranking, creative writing evaluation, and image annotation for LLM training workflows. Native Bahasa Indonesia speaker with professional English proficiency - available for multilingual AI training projects, Indonesian-language content evaluation, and translation QA. Brings 10+ years of professional experience in visual quality control, content review, and data-driven decision making from digital marketing leadership roles, which directly translates to the precision, guideline adherence, and consistency required in large-scale annotation work. Core strengths include: RLHF response rating, creative writing and text quality evaluation, image annotation, content moderation, and Indonesian-language AI training. Education: Bachelor's Degree in Information Systems (STMIK Pratama Indonesia, 2018) and Multimedia Major (SMK Tri Dharma II Bogor, 2017). Available for remote, long-term annotation projects across any major platform including Label Studio, CVAT, Scale Studio, Labelbox, and Encord.

Labeling Experience

Freelance AI Data Annotator – Image Annotation & Visual QA, Outlier (Scale AI)

Scale AIImageClassification

I annotate and evaluate images for visual quality, content accuracy, and guideline compliance as part of multimodal AI model training workflows on Outlier. My work involves assigning class labels, identifying quality issues, and providing structured visual QA feedback that feeds into downstream model training. I maintain consistency across large batches by strictly following evolving annotation protocols and platform requirements. • Label images with accurate class assignments based on detailed project taxonomies • Evaluate visual quality against project-specific standards (composition, clarity, relevance, safety) • Flag edge cases and ambiguities for reviewer clarification to protect dataset integrity • Contribute to large-scale, high-throughput annotation pipelines for multimodal LLM training • Work independently across global time zones with reliable daily throughput

I annotate and evaluate images for visual quality, content accuracy, and guideline compliance as part of multimodal AI model training workflows on Outlier. My work involves assigning class labels, identifying quality issues, and providing structured visual QA feedback that feeds into downstream model training. I maintain consistency across large batches by strictly following evolving annotation protocols and platform requirements. • Label images with accurate class assignments based on detailed project taxonomies • Evaluate visual quality against project-specific standards (composition, clarity, relevance, safety) • Flag edge cases and ambiguities for reviewer clarification to protect dataset integrity • Contribute to large-scale, high-throughput annotation pipelines for multimodal LLM training • Work independently across global time zones with reliable daily throughput

2026 - Present

Freelance AI Data Annotator – Creative Writing & Text Quality Evaluator, Outlier (Scale AI)

Scale AITextRLHF

I rate and evaluate creative writing and text quality outputs from LLMs on Outlier, focusing on tone, logical consistency, factual accuracy, coherence, and prompt adherence. My evaluations use structured task guides and detailed rubrics that feed directly into model fine-tuning and RLHF training pipelines. • Rate individual LLM responses across multiple quality dimensions using detailed task-specific rubrics • Compare and rank alternative responses for preference-based training data • Identify and flag issues in tone, logical flow, factual accuracy, and prompt alignment • Provide structured written feedback to improve model behavior on edge cases • Maintain consistent judgment standards across long task queues

I rate and evaluate creative writing and text quality outputs from LLMs on Outlier, focusing on tone, logical consistency, factual accuracy, coherence, and prompt adherence. My evaluations use structured task guides and detailed rubrics that feed directly into model fine-tuning and RLHF training pipelines. • Rate individual LLM responses across multiple quality dimensions using detailed task-specific rubrics • Compare and rank alternative responses for preference-based training data • Identify and flag issues in tone, logical flow, factual accuracy, and prompt alignment • Provide structured written feedback to improve model behavior on edge cases • Maintain consistent judgment standards across long task queues

2026 - Present

Freelance AI Data Annotator – RLHF Evaluator, Outlier (Scale AI)

Scale AITextRLHF

I perform RLHF preference ranking tasks, evaluating and ranking LLM responses in both Bahasa Indonesia and English. I assess helpfulness, accuracy, coherence, and safety of AI-generated outputs based on detailed task-specific rubrics. My work requires strict adherence to evolving annotation guidelines and consistent high throughput across daily task batches. • Compare and rank multiple LLM responses for preference-based model training • Apply native Bahasa Indonesia and professional English proficiency across multilingual tasks • Evaluate responses against detailed rubrics covering helpfulness, accuracy, coherence, and safety • Flag edge cases and provide structured feedback to improve model alignment • Maintain consistent judgment standards across long task queues and evolving guidelines

I perform RLHF preference ranking tasks, evaluating and ranking LLM responses in both Bahasa Indonesia and English. I assess helpfulness, accuracy, coherence, and safety of AI-generated outputs based on detailed task-specific rubrics. My work requires strict adherence to evolving annotation guidelines and consistent high throughput across daily task batches. • Compare and rank multiple LLM responses for preference-based model training • Apply native Bahasa Indonesia and professional English proficiency across multilingual tasks • Evaluate responses against detailed rubrics covering helpfulness, accuracy, coherence, and safety • Flag edge cases and provide structured feedback to improve model alignment • Maintain consistent judgment standards across long task queues and evolving guidelines

2026 - Present

Education

Work History