I’m a senior full-stack software engineer with 25+ years of experience building software, currently focused on the intersection of AI agents, graph systems, evaluation, provenance, and automation.

My main interest is not wrapping an API around a chatbot. I build systems where AI has to operate inside explicit architecture: workflows, state machines, graphs, evaluation loops, provenance trails, deterministic tooling, and human oversight.

Provenance infrastructure for AI-assisted research and engineering.

EvidEngine turns source material, claims, transformations, and supporting evidence into an inspectable provenance graph. The goal is simple: AI outputs should come with receipts.

Current areas of research and development include:

• causal and temporal provenance graphs
• entity and concept resolution
• deterministic evidence extraction
• signed attestations
• reproducibility
• provenance-aware evaluation
• graph-based retrieval and inspection

Website: https://evidengine.com

Evolutionary optimization for AI agent harnesses.

Whetstone explores using genetic and evolutionary algorithms to improve agent systems by evolving components such as prompts, policies, routing strategies, tools, context strategies, and workflow structure.

A central question behind the project:

How do you evaluate a component individually without optimizing it at the expense of the system it belongs to?

Infrastructure for designing, testing, benchmarking, and evolving agents.

The project explores reproducible agent specifications, evaluation environments, budget-constrained execution, comparative benchmarking, and systematic experimentation with agent architectures.

Agentic automation for go-to-market operations.

An experimental virtual Chief Marketing Officer (CMO) system for turning business context into structured audiences, campaigns, content, workflows, and measurable actions.

Automated Job Application System.

A graph-driven system for matching candidates and companies, evaluating fit, tailoring application materials, and automating the repetitive parts of the job-search process without pretending every opportunity deserves equal effort.

A recurring theme across my work is treating graphs as an engineering tool rather than a fashionable data structure.

I work with and research:

• knowledge graphs
• property graphs
• provenance graphs
• causal graphs
• execution graphs
• agent workflow graphs
• dependency graphs
• semantic graphs
• ontologies and taxonomies
• graph-based retrieval
• graph evaluation
• evolving graph structures with genetic algorithms

I’m also developing Graph Patterns, an educational project about when graph representations actually improve a system—and when a simpler architecture is the better choice.

My current work focuses heavily on engineering AI systems that are:

• inspectable rather than opaque
• evaluated rather than assumed to work
• bounded rather than endlessly autonomous
• stateful when state actually matters
• deterministic where possible
• provenance-aware
• observable at the component and system level
• designed for failure, recovery, and human intervention

I regularly work with multi-step and graph-based agent architectures using tools such as LangGraph and LangChain, but I do not assume every AI application needs an agent—or every agent needs a graph.

Languages

TypeScript · JavaScript · Python · PHP · SQL

Frontend

React · Vue · Angular · Remix · SvelteKit · Tailwind CSS · D3 · Three.js

Backend

Node.js · Express · Django · Laravel · Symfony · Lumen · CodeIgniter

Data

PostgreSQL · MySQL · MongoDB · Redis · Neo4j

AI / Agent Systems

LangGraph · LangChain · LLM evaluation · RAG · tool-use systems · multi-agent workflows · genetic algorithms

Infrastructure

Docker · Linux · AWS · DigitalOcean · GitHub Actions

I’m particularly interested in problems involving:

• agent architecture
• AI evaluation and benchmarking
• context engineering
• graph engineering
• AI provenance and auditability
• evolutionary computation
• human-in-the-loop automation
• developer tooling
• data visualization
• replacing repetitive knowledge work with reliable software

A lot of my work comes back to the same principle:

Make the system explicit enough that you can inspect it, measure it, improve it, and trust it.

I’m building, researching, and looking for opportunities where I can apply this work to real systems—particularly senior software engineering, AI engineering, agentic systems, graph engineering, and technical product development.

I’m also interested in collaborating with people working on:

• AI infrastructure
• agent evaluation
• provenance
• knowledge graphs
• evolutionary AI
• graph-based systems
• automation products

If you're working on something in that neighborhood, open an issue on one of my projects or reach out through GitHub.