Building with intention, curiosity, and a quiet fire to explore the unknown.
I βm Lannon Khau, a full-stack AI engineer who builds production systems that think, learn, and ship. I specialize in Python, cloud infrastructure, and AI-powered workflows β from document intelligence pipelines to RAG-driven legal tools.
Currently at Sweet James, I engineer end-to-end AI systems spanning Azure Document Intelligence, OCR pipelines, workflow automation with n8n, and conversational RAG agents β all deployed to production and used daily by legal professionals.
My stack spans Flask, AWS, Azure, Supabase, and n8n workflow automation. I architect document intelligence pipelines, build citation verification systems, and design AI agents that generate structured legal documents β turning complex, manual processes into streamlined, production-ready tools.
βI build AI-powered systems that turn raw data into insight, clarity, and production-grade tools.β
2025β2026
Engineering production AI systems for a legal tech firm. Built an AI check validation system with Azure Document Intelligence. Created SJ Chatter Box, a RAG-powered demand letter generator with citation verification and Word export. Architected n8n workflow automation for Salesforce syncing and multi-container orchestration.
2025
Completed Cloudberry AWS Bootcamp and advanced GenAI systems training. Submitted Data Forge to the Microsoft Reactor Hackathon. Deployed this portfolio on EC2 with S3, RDS, Secrets Manager, and Nginx with SSL.
2024
Launched Data Forge, an AI-powered EDA tool that turns messy financial data into insights in seconds, delivering sub-2 s query performance across 1.7 M+ transactions.
2023
Produced 2,500 + training prompts for LLM fine-tuning, hitting a 93 % quality benchmark and building automation that cut manual QA by 40 %.
2022
Led the build of a K-12 learning platform (Flask + React + AWS) that added ML-driven feedback and personalized pathsβraising certification pass rates to 87 % and trimming grading time by 50 %.
2021
Earned my B.S. in Information Systems from Cal State Fullerton. Interned at Rhombus Systems, contributing to Python backend services for smart-security platforms and sharpening my passion for real-time data and APIs.
Dive into my featured projects below. Click on any project to see details, the tech stack, and the story behind the build. Every project reflects my hands-on work in AI, automation, and data-driven apps.
These are the production AI agents I architected across my projects at Sweet James. Each one is a specialized component that handles a specific task in a larger intelligent pipeline β from streaming demand letters to verifying citations against source PDFs. Click on any agent to learn how it works under the hood.
Stream Multi-Section Legal Documents
Stream Multi-Section Legal Documents
Map Citations to Source PDFs
Map Citations to Source PDFs
Auto-Fill Placeholder Tags in Drafts
Auto-Fill Placeholder Tags in Drafts
Semantic Search Across Case Documents
Semantic Search Across Case Documents
Classify Legal Documents by Type
Classify Legal Documents by Type
Extract Financial Data from Checks
Extract Financial Data from Checks
Distribute OCR Across Azure Instances
Distribute OCR Across Azure Instances
Extract Structured Data from Salesforce
Extract Structured Data from Salesforce
Individual agents are powerful, but teams of agents solve complex business problems. Here's how I orchestrate specialized AI agents into intelligent workflows that deliver results.
An interactive AI Copilot for business intelligence. Orchestrates natural language SQL agents, data validation, and visualization modules to let you query, analyze, and visualize any business database with plain Englishβno manual SQL required.
SQL Agent
β
Validation
β
Charting
β
Visual
β
An interactive AI Copilot for business intelligence. Orchestrates natural language SQL agents, data validation, and visualization modules to let you query, analyze, and visualize any business database with plain Englishβno manual SQL required.
SQL Agent
β
Validation
β
Charting
β
Visual
β
An interactive AI Copilot for business intelligence. Orchestrates natural language SQL agents, data validation, and visualization modules to let you query, analyze, and visualize any business database with plain Englishβno manual SQL required.
SQL Agent
β
Validation
β
Charting
β
Visual
β