Kushagra Agarwal

Key Skills

Software

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Document

Image

Text

Top Label Types

Action Recognition

Classification

Data Collection

Evaluation Rating

Object Detection

Point Key Point

Question Answering

Relationship

Text Generation

Tracking

Freelancer Overview

I have a strong background in scientific research and engineering, with extensive experience in data collection, analysis, and interpretation across electrochemical and materials science domains. My work has involved designing and implementing rigorous experimental protocols, ensuring data integrity, and supervising teams to accurately capture, label, and validate complex datasets using advanced instrumentation such as XRD, SEM, EDX, TEM, and XRF. I am proficient in scientific software tools like JMP, Matlab, and OriginLab, which I have used to annotate, process, and visualize large volumes of experimental data for R&D projects. My ability to translate raw experimental outputs into actionable insights and my attention to detail in documenting and reporting results equip me to contribute effectively to data labeling and AI training data initiatives, particularly in technical and scientific environments.

Labeling Experience

AI Trainer Expert

Scale AIImageQuestion AnsweringObject Detection

- Evaluated and refined AI model responses for reasoning quality, ethical alignment, factual accuracy, and instruction-following. - Designed and reviewed high-precision prompts to test nuanced concepts such as loyalty, ethics, ambiguity, and decision-making under constraints. - Performed comparative analysis of multiple model outputs, identifying subtle inconsistencies, bias, and logical gaps.

- Evaluated and refined AI model responses for reasoning quality, ethical alignment, factual accuracy, and instruction-following. - Designed and reviewed high-precision prompts to test nuanced concepts such as loyalty, ethics, ambiguity, and decision-making under constraints. - Performed comparative analysis of multiple model outputs, identifying subtle inconsistencies, bias, and logical gaps.

2025 - 2025

Education

Work History