Skip to content
Paul Luckey

Paul Luckey

Applied AI Architect

Austin, Texas

I build AI systems that make people smarter at their jobs. Not chatbots — ambient intelligence that's contextual, real-time, and designed around how humans actually think. I studied psychology, spent 10 years building enterprise systems, and now I design AI experiences at the intersection of cognition and engineering. I've shipped to millions of users, built with zero budget, and can go from problem to production-grade experience faster than most teams.

Impact

309kemployees using LLM semantic search I built with zero budget — reduced the cognitive load of finding the right information by 40%
100+specialists using RAG agent I built the retrieval infrastructure for — chunking, embeddings, cosine similarity, DBT pipelines
500+companies on the risk platform where I now design ambient AI — incident intelligence, real-time NER, pattern detection

What I Build

Ambient AI Experiences

Intelligence that surrounds the human decision-maker — surfacing patterns, risks, and context they wouldn't see alone

RAG & Knowledge Retrieval

Production agents with vector search, embedding pipelines, pgvector, Snowflake Cortex — built for how people actually search

Real-Time AI Pipelines

Streaming inference, entity extraction, risk classification — orchestrated into information-dense dashboards

Multi-Agent Reasoning

Structured multi-perspective deliberation, expert panel orchestration, consensus synthesis

Rapid Prototyping

Problem to production-grade experience in days — Cursor, Claude Code, Kiro, Vercel AI SDK

AI Engineering Enablement

Training teams on LLM-assisted development, prompt engineering, AI-native workflows

Experience

Product Architect/Origami Risk

2025 –
  • Designing ambient AI for enterprise risk management — incident intelligence that augments safety reviewers with real-time NER, similarity search, and pattern detection
  • Rapid prototyping production AI features using Cursor, Claude Code, and Kiro — concept to working system in days
  • Training engineering teams on LLM-assisted development practices and emerging AI techniques

Solution Architect/Insperity

2023 – 2025
  • Built LLM semantic search with zero budget using Snowflake Cortex — transformed how 309k employees find information, 40% drop in search abandonment
  • Built the RAG infrastructure — DBT pipelines, document chunking, embedding generation, cosine similarity retrieval — powering an agent used by 100+ specialists daily
  • Designed API Gateway (YARP) with blue-green deployments — 99.9% uptime across 500k+ daily transactions
  • Built data pipeline architecture with DBT for enterprise-scale HR data transformation

Senior Software Engineer/Insperity

2022 – 2023
  • .NET microservices at enterprise scale — schema design, RESTful APIs, query performance optimization

Software Engineer/Insperity

2021 – 2022

Software Engineer/Enterprise Products

2019 – 2021

Senior Consultant/Gimmal

2017 – 2018

Consultant/Hitachi Vantara

2014 – 2017

Stack

AI/ML
BedrockTitan EmbeddingsComprehendSnowflake CortexRAGpgvectorPrompt EngineeringStreaming InferenceMulti-Agent OrchestrationVector Search
Languages
.NETC#TypeScriptPythonSQL
Frontend
ReactNext.jsTailwindVercel AI SDK
Infrastructure
AWSLambdaS3DockerVercelSupabasePostgreSQLSQL ServerRedisMongoDBRabbitMQSnowflakeDBT
Patterns
MicroservicesEvent-DrivenAPI GatewayBlue-GreenCI/CD
DevOps
Azure DevOpsOctopus DeployDockerGitHub Actions
AI Tooling
Claude CodeCursorKiroCopilot

Education

B.S. PsychologyUniversity of Houston, 2012

Trained in human cognition and behavior — the foundation for designing AI systems around how people actually think and decide.