Joshua Dell

Proof of work

Each one is live, open source, and breakable on purpose, with the architecture documented down to the decision history. Short on time? Start with NextAurora and kill its payment service.

NextAurora

An event-driven microservices store with a kill switch: kill a service live and watch the system not lose a single message.

.NET 10WolverineRabbitMQEF CoreReact 19Keycloak

San Juan Watershed Riparian Watch

Basin-scale riparian monitoring from satellite data: a STAC ETL over Sentinel and Landsat, geospatial ML, and a map-based decision-support UI.

PythonSTACGeospatial ML.NET 10PostGISReact 18

Quartzose

A programmatic AI platform for enterprise documents: grounded, cited answers over contracts and filings, model-agnostic by design, with a per-question dial between engineer-authored and model-authored control flow.

PythonFastAPIHybrid retrievalCorrective RAGQdrantvLLM

okl — the Observed Knowledge Ledger

Load your team's rules, ADRs and standards, hold your coding agent to them, and keep learning as you build: proven by checks, flagged when the code they cover changes.

PythonSQLite / PostgresFastAPIMCPClaude CodeGitHub Actions

What I build

I own the whole arc, from requirements and architecture through deployment, production support and iteration. I build RAG and document-intelligence systems, agentic AI workflows, and geospatial platforms, and I use AI tooling daily to rethink how engineering gets done.

• Full-stack production ownership. End-to-end product ownership: requirements, architecture, shipping, iteration. Production APIs in .NET and Python; Azure infrastructure; observability with structured logging and correlation IDs as defaults.
• Data pipelines & ML integration. 20+ production ETL/ELT pipelines with real-time integrations; document processing at scale; RAG architecture, embeddings, and agentic AI workflows; event-driven distributed systems.
• Geospatial & environmental systems. Satellite-data pipelines (Sentinel, Landsat, STAC), geospatial ML with spatial cross-validation, and map-based decision-support UX, built on a decade of GIS engineering and an environmental-studies foundation.
• Hands-on technical leadership. Leads through implementation, not just design docs; mentors on system design, code quality, and testing discipline; champions AI-assisted engineering that accelerates without sacrificing quality.