Work
Systems built to last.
A selection of projects across AI architecture, cloud infrastructure, and automation. The names are generic where they need to be; the problems were not.
-
QualProcessor
A qualitative analysis pipeline that replaced weeks of manual researcher work with under 5 minutes of automated processing. Uses map-reduce tagging, comment classification, statistical validation, and RAG-generated summaries with citations and numbers.
- Python
- LangChain
- RAG
- pgvector
- AWS ECS
-
SurveyBuilder
An agentic survey building system built with LangGraph. First agentic system on the platform, featuring persistent memory and full AWS ECS deployment.
- Python
- LangGraph
- PostgreSQL
- AWS ECS
-
QuantProcessor
Automated quantitative survey analysis pipeline that generates crosstab analysis and produces AI-generated summaries with insights on top of the results.
- Python
- LangChain
- RAG
- pgvector
-
DataAnalyzer
Automated quality checker for survey responses that flags problematic entries: speedsters, duplicate IPs, crossliners, gibberish, and other data quality issues.
- Python
- LangChain
- AWS Bedrock
-
AI Microservice Template
A scaffolded project template enabling engineers to go from zero to a deployable AI microservice in hours. Includes convention files, scaffolding skill, and feature breakdown skill designed for Claude Code-assisted development. Three-container architecture: API, worker, and agent deployable to AWS Bedrock Agent Core.
- Python
- LangChain
- pgvector
- SQLAlchemy
- AWS ECS
- Terraform
- Claude Code
-
Bedrock Platform Migration
Migrated the AI platform from OpenAI to AWS Bedrock enabling multi-LLM flexibility. Includes agentic memory, prompt management, and guardrails via Bedrock Agent Core runtime.
- AWS Bedrock
- Strands Agents
- LangGraph
- Bedrock Agent Core
Working on something that needs a steady hand?
Let's talk →