Gugan Loganathan

Key Skills

Software

AWS SageMaker

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Image

Top Task Types

Audio Recording

Computer Programming/Coding

Data Collection

Prompt + Response Writing (SFT)

Text Summarization

Freelancer Overview

I bring strong experience in AI training data, evaluation, and data labeling, with a focus on large language models (LLMs), speech, and multimodal systems. At Amazon, I’ve worked hands-on with annotation workflows, maintaining over 98%+ QA accuracy in speech/text labeling tasks, while also contributing to the development of internal tools that improve efficiency and quality. My expertise includes guideline-driven annotation, quality assurance, LLM evaluation, and multimodal content description, ensuring that datasets are both precise and impactful for model training. Beyond industry work, I’ve built and deployed real-world projects such as AutoFlowLite, a real-time LLM assistant deployed on Hugging Face, and the DA Workflow Check-In Dashboard, a Streamlit + AWS QuickSight solution for live productivity tracking. I’m also the first author of two AI research papers under review at IEEE and NeurIPS, focusing on disaster response and healthcare applications. This combination of practical industry experience, technical project ownership, and research contributions sets me apart as someone who can deliver high-quality, reliable AI training data across domains.

Labeling Experience

Speech & Text Annotation for Conversational AI Systems

Aws SagemakerTextEntity Ner ClassificationEvaluation Rating

As part of my role at Amazon, I worked on large-scale annotation of speech and text prompts for conversational AI and LLM training. Tasks included precise text categorization, entity recognition, prompt/response evaluation, and audio recordings for speech model training. I consistently maintained >98% QA accuracy while adhering to strict linguistic and stylistic guidelines. The project involved labeling and validating thousands of data points weekly, ensuring high-quality datasets for fine-tuning and evaluation of next-generation AI systems.

As part of my role at Amazon, I worked on large-scale annotation of speech and text prompts for conversational AI and LLM training. Tasks included precise text categorization, entity recognition, prompt/response evaluation, and audio recordings for speech model training. I consistently maintained >98% QA accuracy while adhering to strict linguistic and stylistic guidelines. The project involved labeling and validating thousands of data points weekly, ensuring high-quality datasets for fine-tuning and evaluation of next-generation AI systems.

2025

Education

Work History