Sunny K.

Key Skills

Software

Don't disclose

Snorkel AI

Top Subject Matter

Reinforcement Learning from Human Feedback for LLMs

AI model training data preparation and SFT evaluation

Top Data Types

Text

Video

Document

Top Task Types

RLHF

Prompt + Response Writing (SFT)

Evaluation/Rating

Computer Programming/Coding

Freelancer Overview

Freelance RLHF Engineer, Air Dawg Remote (Nov 2025 – Present). Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose. Education includes Bachelor of Technology, Indian Institute of Information Technology (IIIT), Surat (2024). AI-training focus includes data types such as Text and labeling workflows including RLHF , Prompt and Response Writing (SFT).

Labeling Experience

Freelance RLHF Engineer, Air Dawg Remote (Nov 2025 – Present)

Don't discloseTextRLHF

Performed RLHF-based evaluation and ranking of large language model outputs to improve response quality and alignment. Annotated model outputs for reasoning quality, factual accuracy, and safety constraints to guide preference learning. Supported prompt engineering and red-teaming workflows aimed at reducing algorithmic bias and alignment errors. • Evaluated LLM responses for reasoning, factual accuracy, and safety • Ranked and annotated outputs for RLHF optimization • Participated in prompt engineering iterations • Contributed to red-teaming pipeline improvements

2025 - Present

Freelance AI Model Trainer, Xelron Remote (May 2025 – Oct 2025)

Don't discloseTextPrompt + Response Writing (SFT)

Curated and preprocessed high-quality training datasets used for fine-tuning multi-modal AI models. Wrote gold-standard, structured responses to perform supervised fine-tuning (SFT) checks for complex technical queries. Conducted error analysis on model predictions to identify performance bottlenecks and inform remediation. • Curated and preprocessed training datasets for fine-tuning • Authored logical gold-standard responses for SFT checks • Analyzed prediction errors to locate weaknesses • Supported dataset and response quality improvements

2025 - 2025

Education

Work History