Dhanalakshmi B

Key Skills

Software

AWS SageMaker

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Document

Image

Text

Video

Top Label Types

Action Recognition

Bounding Box

Classification

Evaluation Rating

Prompt Response Writing SFT

Segmentation

Text Generation

Text Summarization

Freelancer Overview

I am an experienced AI Training and Data Annotation Specialist with a strong background across image, video, text, audio, and search-based datasets. My work at Amazon involved high-precision annotation, action and object segmentation, query evaluation, and consistent QA auditing. I am skilled in evaluating LLM responses, identifying errors, refining guidelines, and maintaining high accuracy across diverse AI projects.

Labeling Experience

LLM Safety & Red Team Evaluation - Amazon internal AI models

Aws SagemakerTextEvaluation RatingRed Teaming

I worked on evaluating LLM responses for safety, factual correctness, user-intent alignment, and guideline compliance. My work followed a red-teaming style approach where I tested model behavior with diverse prompts, identified unsafe or biased outputs, and documented failure cases. I performed detailed audits, flagged harmful or policy-violating content, clarified ambiguous cases, and provided structured feedback to improve model safety and robustness. This included checking edge cases, applying strict labeling rules, and ensuring consistency across multilingual datasets. I also reviewed other annotators’ work to maintain high accuracy standards.

I worked on evaluating LLM responses for safety, factual correctness, user-intent alignment, and guideline compliance. My work followed a red-teaming style approach where I tested model behavior with diverse prompts, identified unsafe or biased outputs, and documented failure cases. I performed detailed audits, flagged harmful or policy-violating content, clarified ambiguous cases, and provided structured feedback to improve model safety and robustness. This included checking edge cases, applying strict labeling rules, and ensuring consistency across multilingual datasets. I also reviewed other annotators’ work to maintain high accuracy standards.

2024

Data Associate

Aws SagemakerImageBounding BoxPolygon

I worked on high precision image and video annotation for Amazon's Consumer Robotics team. My task included creating bounding boxes, polygons and pixel level segmentation for object detection, object tracking, timestamp alignment, and scene understanding. I handled complex visual datasets, resolved ambiguous cases, performed QA checks, and maintained annotation accuracy above 95-98% while adhering to strict guidelines.

I worked on high precision image and video annotation for Amazon's Consumer Robotics team. My task included creating bounding boxes, polygons and pixel level segmentation for object detection, object tracking, timestamp alignment, and scene understanding. I handled complex visual datasets, resolved ambiguous cases, performed QA checks, and maintained annotation accuracy above 95-98% while adhering to strict guidelines.

2023

Amazon – Machine Learning Data Associate

Aws SagemakerTextBounding BoxPolygon

I worked on high-precision video and image annotation for Amazon’s Consumer Robotics team. My responsibilities included frame-by-frame action identification, object tracking, pixel-level segmentation, and creating high-quality labels for training computer vision models. I also performed QA audits, corrected inconsistencies, clarified guideline ambiguities, and maintained accuracy levels above 95%. The project involved strict quality standards, timestamp alignment, scenario validation, and reviewing edge cases to ensure consistent model training data. I also supported the team by training new associates, assigning tasks through the internal tool, and helping maintain overall annotation quality.

I worked on high-precision video and image annotation for Amazon’s Consumer Robotics team. My responsibilities included frame-by-frame action identification, object tracking, pixel-level segmentation, and creating high-quality labels for training computer vision models. I also performed QA audits, corrected inconsistencies, clarified guideline ambiguities, and maintained accuracy levels above 95%. The project involved strict quality standards, timestamp alignment, scenario validation, and reviewing edge cases to ensure consistent model training data. I also supported the team by training new associates, assigning tasks through the internal tool, and helping maintain overall annotation quality.

2023

Education

Work History