Hari Raja Prabhu

Key Skills

Software

Google Cloud Vertex AI

Labelbox

Mercor

Scale AI

V7 Labs

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Entity Ner Classification

Text Generation

RLHF

Evaluation Rating

Prompt Response Writing SFT

Freelancer Overview

I have practical experience in AI training data preparation, annotation, and quality validation across text and structured datasets. My work includes text classification, sentiment labeling, entity tagging (NER), intent annotation, data categorization, and reviewing model-generated outputs for accuracy and bias. I strictly follow annotation guidelines, maintain high inter-annotator agreement standards, and perform detailed quality checks to ensure clean, reliable training data. With a strong software engineering background, I understand the end-to-end ML lifecycle, including how high-quality labeled data impacts model training, evaluation, and performance optimization. I am experienced in identifying edge cases, resolving ambiguous samples, documenting annotation decisions, and maintaining consistency at scale. My analytical thinking, attention to detail, and ability to work with structured workflows make me well-suited for AI training, LLM evaluation, and data labeling projects.

Labeling Experience

LLM Response Evaluation & Text Annotation for AI Model Improvement

Scale AITextEntity Ner ClassificationText Generation

Worked on large-scale text annotation and evaluation tasks to improve LLM performance and alignment. Responsibilities included rating AI-generated responses based on relevance, accuracy, coherence, safety, and instruction-following. Performed comparative ranking of multiple model outputs and provided structured feedback to enhance reinforcement learning from human feedback (RLHF) pipelines. Contributed to supervised fine-tuning (SFT) by writing high-quality prompt-response pairs, correcting model hallucinations, identifying edge cases, and flagging bias or unsafe outputs. Maintained strict adherence to annotation guidelines, ensured consistency across tasks, and achieved high agreement scores during quality audits. Regularly collaborated with QA teams to refine labeling standards and improve dataset reliability.

Worked on large-scale text annotation and evaluation tasks to improve LLM performance and alignment. Responsibilities included rating AI-generated responses based on relevance, accuracy, coherence, safety, and instruction-following. Performed comparative ranking of multiple model outputs and provided structured feedback to enhance reinforcement learning from human feedback (RLHF) pipelines. Contributed to supervised fine-tuning (SFT) by writing high-quality prompt-response pairs, correcting model hallucinations, identifying edge cases, and flagging bias or unsafe outputs. Maintained strict adherence to annotation guidelines, ensured consistency across tasks, and achieved high agreement scores during quality audits. Regularly collaborated with QA teams to refine labeling standards and improve dataset reliability.

2024

Education

Work History