Anand Kumar

Key Skills

Software

Argilla

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Top Label Types

Entity Ner Classification

Question Answering

Text Summarization

Translation Localization

Freelancer Overview

As a freelance AI trainer with a strong foundation in engineering and data handling, I specialize in training and evaluating large language models (LLMs), curating high-quality datasets, and refining model outputs for improved performance. I have hands-on experience with text classification, prompt generation, sentiment analysis, and basic code annotation. My academic background in BTech Biotechnology adds a unique analytical approach, allowing me to work on scientific content, education-related datasets, and language-based AI systems with precision. I’ve contributed to projects involving AI model alignment, chatbot response evaluation, and instruction tuning. My familiarity with tools like AWS SageMaker, Argilla, and Appen, combined with strong communication and attention to detail, makes me a valuable asset in any AI training pipeline. I'm passionate about building ethical, reliable AI systems and continually expanding my capabilities in both supervised labeling and human-in-the-loop model training.

Labeling Experience

Prompt Evaluation & Response Ranking for Large Language Model

AppenTextText GenerationPrompt Response Writing SFT

Worked as part of a global team to evaluate and improve responses generated by large language models. Tasks involved comparing multiple AI-generated answers, ranking them based on helpfulness, coherence, and accuracy, and providing written feedback to guide model tuning. Also created and labeled user prompts to improve real-world instruction-following capability.

Worked as part of a global team to evaluate and improve responses generated by large language models. Tasks involved comparing multiple AI-generated answers, ranking them based on helpfulness, coherence, and accuracy, and providing written feedback to guide model tuning. Also created and labeled user prompts to improve real-world instruction-following capability.

2024

Education

Work History