Divyam G.

Key Skills

Software

Scale AI

Remotasks

Appen

Clickworker

Data Annotation Tech

Mercor

Micro1

Mindrift

Top Subject Matter

Artificial Intelligence & LLM Development

Finance & Banking

Speech & Language / Multilingual NLP

Top Data Types

Document

Audio

Image

Top Task Types

RLHF

Evaluation/Rating

Transcription

Text Summarization

Data Collection

Object Detection

Entity (NER) Classification

Classification

Freelancer Overview

Experienced AI data labeler and trainer with hands-on work across multiple platforms including Outlier AI, Scaler AI (Dreamliner), Crowdgen, and Appen. Skilled in RLHF response evaluation, prompt writing, preference ranking, and assessing LLM outputs across coding, reasoning, and creative writing tasks. Additionally experienced in evaluating AI agent behavior across real-world software tools such as MS Word, Zendesk, Gmail, and Stripe using structured task rubrics. Proficient in Hindi audio transcription for multilingual speech and NLP datasets, with native-level Hindi fluency. Known for delivering consistent, high-quality work while adhering to strict platform guidelines and quality benchmarks. Passionate about contributing to safe, accurate, and capable AI systems through precise human feedback.

Labeling Experience

AI Task Evaluator – Real-World Software Workflows

TextEvaluation/Rating

Evaluated AI-generated task completions on Scaler AI's Dreamliner platform, assessing whether outputs correctly followed structured task rubrics across real-world software environments including MS Word, Zendesk, Gmail, and Stripe. Reviewed multi-step workflows for accuracy, logical correctness, and adherence to instructions. Provided detailed pass/fail judgments and qualitative feedback to help train AI agents on tool-use and task execution.

2025 - Present

Hindi Transcription & Localization Specialist

AudioTranscription

Performed Hindi audio transcription tasks on Crowdgen and Appen platforms, accurately converting spoken Hindi content into clean, structured text for use in AI speech and language model training datasets. Followed strict transcription guidelines including handling of dialects, fillers, overlapping speech, and punctuation conventions. Maintained high accuracy rates while meeting platform quality benchmarks and delivery timelines.

2025 - 2025

AI Trainer & Response Evaluator

TextRLHF

Worked as an AI trainer on Outlier AI's platform, evaluating and ranking large language model responses for quality, accuracy, coherence, and safety. Tasks included writing and refining prompts across domains such as coding, reasoning, math, and creative writing. Provided detailed comparative feedback to support RLHF pipelines, ensuring outputs aligned with helpfulness and harmlessness guidelines. Maintained high accuracy and consistency across hundreds of tasks.

2024 - 2025

Education

Work History