Khaja R.

Key Skills

Software

Appen

Clickworker

Labelbox

Mercor

Micro1

Mindrift

OneForma

Remotasks

Scale AI

Toloka

Surge AI

Sama

Top Subject Matter

Artificial Intelligence & Machine Learning

Generative AI & Large Language Model (LLM) Evaluation

Software Engineering & Code Intelligence

Top Data Types

Text

Computer Code Programming

Document

Top Task Types

RLHF

Evaluation/Rating

Computer Programming/Coding

Prompt + Response Writing (SFT)

Function Calling

Fine-tuning

Classification

Data Collection

Transcription

Segmentation

Freelancer Overview

I have hands-on experience in AI training data and data labeling through my work as an AI Trainer and Reviewer on large-scale, Google-aligned RLHF projects. I have completed 1500+ high-quality training and evaluation tasks, focusing on improving LLM outputs through response ranking, prompt evaluation, and structured human feedback. My work involved assessing code correctness, logical reasoning, instruction adherence, and edge case handling, ensuring models produce accurate, reliable, and production-ready responses. I was promoted to a reviewer role, where I audited and improved over 1000 tasks, maintaining high quality standards across training pipelines. What sets me apart is my strong blend of software engineering and applied AI expertise. I have contributed to real-world AI systems by designing bug fixes, feature implementations, and production-grade code patches, particularly in projects simulating real developer workflows. With proficiency in Python and experience in prompt engineering, LLM evaluation, and QA processes, I bring a deep understanding of both model behavior and practical deployment requirements. My ability to combine technical rigor with human-centered evaluation allows me to significantly enhance model alignment, consistency, and real-world usability.

Labeling Experience

Senior AI Trainer & Reviewer (RLHF & LLM Code Evaluation) - Meridial (Invisible Technologies)

Computer Code ProgrammingRLHF

AI Trainer and Reviewer on Google-aligned RLHF training systems, focused on improving LLM response quality, reasoning, and reliability through structured human feedback. Evaluated and refined model outputs for correctness, logical reasoning, edge-case handling, and instruction adherence, contributing to robust and production-ready AI systems with strong code intelligence capabilities. Completed 500+ high-quality AI training tasks and was promoted to Reviewer, where I reviewed 1000+ tasks with a focus on code correctness, reasoning accuracy, and consistency. Performed comprehensive QA validation, including edge case analysis, test coverage, and instruction-following checks, while contributing to RLHF pipelines to enhance model alignment, reliability, and response consistency.

2025 - Present

AI Training Data Specialist (RLHF & LLM Evaluation) - Soul AI by Deccan AI

TextRLHF

Worked on an RLHF-based AI training project focused on improving the performance and alignment of Large Language Models (LLMs). The role involved evaluating, ranking, and refining model-generated responses based on accuracy, relevance, logical reasoning, and instruction adherence. Provided structured human feedback to enhance model understanding, reduce hallucinations, and ensure consistency across diverse prompts and real-world scenarios. Contributed to high-quality data annotation and evaluation pipelines by following strict guidelines and maintaining consistency in labeling. Actively identified errors in model outputs, including reasoning gaps and incomplete responses, and suggested improvements to optimize overall model behavior. The project aimed to enhance the reliability, usability, and real-world applicability of AI systems across various domains.

2025 - 2025

Associate Software Engineer at Tech Mahindra

Computer Code ProgrammingComputer Programming/Coding

Worked as an Associate Software Engineer on an enterprise-grade Digital Conveyancing Portal for the Singapore Land Authority, focusing on frontend development using React, JavaScript, and TypeScript. Contributed to building scalable and user-friendly UI components, improving application performance, and ensuring seamless integration with backend microservices through REST APIs. Actively participated in requirement analysis, collaborating with stakeholders and cross-functional teams to deliver high-quality features aligned with business needs. Implemented unit testing using Jest and React Testing Library to enhance test coverage and maintain code quality standards, contributing to improved SonarQube scores. Played a key role in debugging, performance optimization, and code reviews, ensuring a smooth and efficient user experience. Utilized GitLab for version control and CI/CD workflows, supporting the successful release of the project by delivering reliable, maintainable, and production-ready frontend solutions.

2023 - 2025

Education

Work History