Gary Dotzlaw
Senior Software Architect & AI Systems Engineer
I've shipped production software for 100+ companies across 35 years — through every paradigm shift the industry has thrown at us: desktop, client-server, web, cloud, and now agentic AI. That depth means I don't just use these tools. I architect the infrastructure that makes them reliable.
My current focus is the infrastructure layer that makes agentic AI reliable in production: swarm architecture, defense-in-depth security, and a migration methodology that moves legacy codebases into modern AI frameworks. The Bootstrap Framework — an agent swarm that builds agent swarms — is validated across 3 production projects, with each migration completing faster than the last.
When I'm not in the code: fishing, woodworking, and photography.
My Skills
Claude Code Expert
Deep expertise across the full Claude Code ecosystem: agent teams with file ownership boundaries, hook engineering for deterministic control, progressive disclosure skills, slash commands, and the Bootstrap Framework — an agent swarm that builds agent swarms, validated across 3 production migrations.
Agentic AI Security
Applying the OWASP Top 10 for Agentic Applications: 22-pattern prompt injection detection, two-tier scan enforcement, rate limiting circuit breakers, secrets hygiene (SEC-01–05), and a 3-tier trajectory monitoring protocol.
Python Backend Engineering
FastAPI with Pydantic models, asyncio.TaskGroup, centralized PostgreSQL connection pooling, full type annotations. Experience migrating monolithic architectures to modular router-based designs. Modern Python typing throughout.
AI Pipeline Engineering
Multi-step AI processing pipelines with Claude API, OpenAI embeddings, Qdrant vector search (two-stage retrieval), BaseLLMClient abstraction, WebSocket progress reporting, and classified error recovery.
SQL and Data Engineering
Expert-level SQL across MS SQL Server and PostgreSQL: multi-tenant query systems, view-centric architecture, T-SQL dialect specifics, natural language to SQL pipeline design with schema retrieval via vector search.
React and TypeScript Frontend
React 19 with React Query, TypeScript strict mode, shadcn/ui, Metabase SDK embedding, real-time WebSocket integration, Astro 5 + Svelte 5 + Tailwind CSS 3 for static sites. tsc --noEmit enforced on every write.
Hook Engineering and Automation
Building deterministic quality gates: PreToolUse safety gates (72+ destructive command patterns), PostToolUse quality enforcement (ruff, tsc, Biome, artifact validation), Stop hooks (full test suite). additionalContext for self-correction.
Domain Knowledge Extraction
Transforming codebase knowledge into Claude Code skills via progressive disclosure. From metabase-server migration: 7 skills, 1,486 lines covering 7-step SQL workflow, view-centric patterns, LLM client abstraction. 140x token efficiency.
GitHub Copilot Agent Workflows
Designing multi-agent pipelines in GitHub Copilot: 10 specialized agents across research, implementation, review, and domain capture workflows. Self-improving skills via knowledge harvest, domain exploration with resumable frontiers, and JIRA-integrated handoff documentation.
Astro and Modern Static Sites
Astro 5 with Svelte 5, Tailwind CSS 3, Vercel deployment, Pagefind search, Biome linting. Content collections with Zod schema validation, partial hydration for performance, pnpm workspaces.
Servoy Development
Two decades of enterprise business application development with Servoy. Complex form hierarchies, data provider patterns, JavaScript business logic. Current: integrating Servoy with AI — natural language interfaces, automated reports.
RAG & GraphRAG Expert
End-to-end retrieval-augmented generation: vector databases, knowledge graphs, and GraphRAG hybrid architectures. Production systems include a 5-stage pipeline (2,757 links, 2.5s chatbot, $1.50 total) and Text-to-SQL RAG achieving 92–95% accuracy. Published 4-part GraphRAG optimization series with benchmarks.
The Dotzlaw Team
Two skilled engineers building advanced agentic AI projects and research alongside me. They contribute directly to the systems, articles, and tools published on this site.
Building AI-powered data pipelines and full-stack applications at the intersection of machine learning and real-world business problems.
Applying statistical analysis, neural networks, and modern UI to extract insight from complex datasets and build compelling data-driven applications.
Interested in working together?
Get in Touch →