Parth K

Key Skills

Software

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Text

Top Label Types

Computer Programming Coding

Data Collection

Evaluation Rating

Fine Tuning

RLHF

Text Generation

Text Summarization

Freelancer Overview

I have a strong background in computer science with hands-on experience designing and automating data pipelines, validation workflows, and reproducible evaluation systems for AI and software projects. My work has focused on ensuring data quality and correctness through rigorous task formulation, automated assessment frameworks, and data-driven testing, including projects for clients like Stripe, Shopify, Salesforce, and Anthropic. I am proficient in Python, SQL, and data processing libraries such as Pandas, and have built deterministic pipelines for data ingestion, transformation, and verification. My experience collaborating with stakeholders to define clear annotation guidelines and success criteria, along with my expertise in automation and reproducible workflows, enables me to contribute effectively to high-quality data labeling and AI training data initiatives across various domains.

Labeling Experience

Generalist AI Trainer

Scale AITextText Generation

Led evaluation and quality assurance of LLM text generation across advanced mathematics, data science, machine learning, and programming tasks. Architected automated validation and regression frameworks for model workflows using Playwright, Jest, and Vitest, ensuring production-grade reliability. Executed adversarial, multilingual, and stress testing to assess model robustness, safety, and reasoning consistency. Designed scalable annotation and model-evaluation pipelines, reducing regression defects by 40% across iterative releases. Defined and enforced high-precision annotation and QA standards, improving inter-annotator agreement by 15%. Partnered with product, engineering, and research teams to surface complex edge cases early and improve model readiness for deployment.

Led evaluation and quality assurance of LLM text generation across advanced mathematics, data science, machine learning, and programming tasks. Architected automated validation and regression frameworks for model workflows using Playwright, Jest, and Vitest, ensuring production-grade reliability. Executed adversarial, multilingual, and stress testing to assess model robustness, safety, and reasoning consistency. Designed scalable annotation and model-evaluation pipelines, reducing regression defects by 40% across iterative releases. Defined and enforced high-precision annotation and QA standards, improving inter-annotator agreement by 15%. Partnered with product, engineering, and research teams to surface complex edge cases early and improve model readiness for deployment.

2024 - 2025

Education

Work History