Tommy Y.

Key Skills

Software

AWS SageMaker

Top Subject Matter

LLM-enabled operational data summarization

Classification Domain Expertise

and prioritization for delivery workflows

Top Data Types

Text

Document

Audio

Top Task Types

Text Summarization

Data Collection

Segmentation

Evaluation/Rating

Computer Programming/Coding

Function Calling

RLHF

Fine-tuning

Text Generation

Freelancer Overview

Software Development Engineer II - Delivery Experience AI at Amazon. Brings 14+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include AWS SageMaker, Internal, and Proprietary Tooling. Education includes Master of Science in Computer Science, Georgia Institute of Technology (2022) and Bachelor of Science in Computer Science, Rutgers University (2016). AI-training focus includes data types such as Computer Code and Programming and labeling workflows including Text Summarization and Data Collection.

Labeling Experience

Software Development Engineer II - Delivery Experience AI at Amazon

AWS SageMakerText Summarization

Built internal LLM-powered tooling to support operational workflows for Amazon Delivery Experience, using AI to transform raw operational data into structured insights. Designed and implemented LangChain and AWS Bedrock workflows to summarize, classify, and prioritize operational data for downstream decision-making and reporting. Integrated event-driven backend services with AI outputs to enable scalable, automated processing. • Used LangChain and AWS Bedrock for LLM-based processing • Developed Python/FastAPI services handling AI-enabled event workflows • Leveraged DynamoDB, S3, SQS, SNS, and Lambda for data handling • Built React/TypeScript UI for AI-driven insights consumption by non-technical stakeholders

2025 - Present

Software Development Engineer II – Amazon Go at Amazon

Data Collection

Developed annotation visualization and correction workflows to improve workforce annotation accuracy for Amazon Go operations. Created systems enabling parallel annotation task processing across a large worker pool, sharing sessions to increase throughput and labeling quality. Migrated data processing scripts to parallel cloud compute to reduce turnaround time for annotated data. • Built annotation visualization tool for review and correction • Designed 24+ hour parallel annotation task system across 500+ workers • Implemented mechanisms to share annotation sessions for faster processing • Reduced annotated data turnaround time by parallelizing processing in cloud compute

2020 - 2022

Education

Work History