Sanghwa Na

About

Sanghwa Na works in AI engineering, AI automation & agents and ML & data science, based in San Francisco, CA (Amazon Web Services (@aws)). Architecting AI agents, RAG systems, and production ML deployments. Public projects on GitHub include Personal Ai Task Assistant (The Personal AI Task Assistant is an interactive AI tool designed to assist with task automation, project…), Aws Architect Agent (This project uses LangGraph and Amazon Bedrock to generate AWS architecture diagrams based on user queries) and Llama2 Finetuning Deploy (이 프로젝트는 Amazon SageMaker에서 LLama2 Large Language Model (LLM)을 미세 조정하는 방법을 보여줍니다), mostly written in Python, Jupyter and HCL.

Projects 6

• github.com

  Personal Ai Task Assistant

  The Personal AI Task Assistant is an interactive AI tool designed to assist with task automation, project management, and decision-making. This assistant leverages natural language processing to understand user inputs, execute tasks with integrated tools, and provide detailed updates and visualizations of progress.

• github.com

  Aws Architect Agent

  This project uses LangGraph and Amazon Bedrock to generate AWS architecture diagrams based on user queries.

• github.com

  Llama2 Finetuning Deploy

  이 프로젝트는 Amazon SageMaker에서 LLama2 Large Language Model (LLM)을 미세 조정하는 방법을 보여줍니다. Hugging Face의 PEFT (Parameter Efficient Fine-tuning) 기법과 QLoRA (Quantized Low-Rank Adapters)를 사용하여, 대규모 언어 모델을 더 효율적으로 미세 조정할 수 있습니다.

• github.com

  Aim306 Kiro Sagemaker

  AWS Summit LA 2026 (AIM306): drive an open-weight LLM to a deployed, benchmarked Amazon SageMaker AI endpoint through a coding agent, steered by portable SKILL.md contracts.

• github.com

  Opensearch Semantic Cache

  이 프로젝트는 LangChain과 OpenSearch를 사용하여 AI 모델의 응답을 캐싱하는 예제를 제공합니다.

• github.com

  Sagemaker Deploy Llms

  AWS SageMaker를 이용하여 HuggingFace의 Large Language Models(LLMs)을 배포하고, 선택적으로 스트리밍을 지원하기 위해 Lambda 함수와 API Gateway를 설정하는 프로젝트입니다.