Aaron (Ke Yi)

AI Engineering Leader · Cloud-Native Architect · Open-Source Builder

Engineering leader and hands-on builder with 18 years in distributed systems and cloud infrastructure, now focused on LLM-powered systems, agent orchestration, and production-grade AI platforms. I believe in rapid prototyping, open-source-driven innovation, and shipping real AI products — not just demos.

18
Years in Systems
7
Patents (US/EU/CN)
9x
AWS Certified
10+
OSS Projects
$ cat ~/profile.yml
focus:
  - LLM Infrastructure & Agent Orchestration
  - RAG Pipelines & Document Intelligence
  - Multi-modal AI Systems
  - Cloud-Native Architecture (Serverless, Containers)

current_stack:
  languages: [Python, TypeScript, Go]
  infra: [AWS, Cloudflare, Docker, Terraform]
  ai: [Bedrock, Claude, GPT, DeepSeek, Stable Diffusion]

motto: "Ship it. Open-source it. Iterate."

What I've Built

Led 1,000+ solution deployments globally with 100+ production launches. Here are some highlights:

LLM-Powered Document Intelligence

Architected and shipped an IDP system with structured data extraction via LLM + validation layers, RAG-based document indexing with citation, and case association algorithms. Production-scale for legal industry.

GenAI Agent Workflows

Rapid prototyping of agent-based workflows, document reasoning, and summarization systems. Built evaluation frameworks for prompt iteration and model benchmarking.

End-to-End GenAI MVPs

From concept to architecture to production — built AI-native cloud solutions at scale integrating document AI, vector pipelines, and serverless infrastructure.

Open-Source Projects

Core contributor and maintainer of 10+ repositories. I build tools I wish existed.

Speaking & Publications

Create your first API from scratch with OpenAPI and Amazon API Gateway — AWS re:Invent 2024, Builders Session
Building Serverless applications with AWS Terraform — AWS re:Invent 2022, Builders Session
Extension for Stable Diffusion on AWS — AWS Workshop (Public)
Building a Distributed Load Testing Engine Using Lambda Container Images — AWS Blog
Developing a Serverless EFS File Browser — AWS Blog
Building Apache Superset on Amazon ECS Fargate — AWS Blog

Career Arc

From low-level network processors to LLM agents — always at the systems layer.

Director of Engineering
AI-Native Legal Tech · 2025 – Present

Building and scaling LLM-powered document processing, RAG pipelines, and agent workflows for legal industry.

Engineering Manager & Software Architect
Major Cloud Provider · 2019 – 2025

Led teams delivering AI-native cloud solutions globally. 1,000+ deployments, 100+ production launches. Core OSS contributor.

Staff Engineer & Tech Lead
Telecom & Semiconductor Industry · 2008 – 2019

Built carrier-grade distributed systems. Hardware-software co-design, network processors, IP/MPLS protocol stacks, segment routing.

M.S. Control Theory & Engineering
Key University · 2005 – 2008

Patents

7 patents granted across US, EU, and China — spanning network resource optimization, SDN, packet processing, and distributed systems.

Let's Connect