Every complex system starts simple, then grows into a maze. I've spent my career spotting what works, writing it down, and making it repeatable for humans and AI tools.

Codebase Rescues

From Spaghetti to Structure

  • Migrated legacy server-rendered apps to modern TypeScript + SPA/SSR architectures—gradually, without breaking production
  • Untangled monoliths into modular monoliths when microservices were overkill
  • Fixed event-driven architectures that had become distributed monoliths
  • Resolved circular dependencies that made entire modules untestable
  • Eliminated god classes that resisted every attempt at modification
  • Rebuilt APIs that had grown inconsistent through years of patches

The Modernizations

  • VM-based deployments to containerized infrastructure with zero downtime
  • Manual deployment processes to GitOps with IaC and policy checks
  • CSS soup to design systems with reusable components
  • Callback hell to clean async/await patterns
  • Proprietary frameworks to standard ecosystems
  • Shadow IT operations to governed, observable infrastructure

Playbooks I've Built

Architecture Playbooks

  • API versioning strategies that don't break clients
  • Module boundaries that actually stay separated
  • Domain-driven design contexts that map to reality
  • Event schemas that evolve without breaking consumers
  • Authentication patterns that work across microservices
  • Feature flag systems that enable safe deployments

Team Playbooks

  • Code review processes that catch problems without blocking progress
  • Architectural Decision Records that people actually read
  • Onboarding documentation that gets new devs productive fast
  • Pattern libraries that make decisions reusable
  • Testing strategies that give confidence without slowing shipping

Systems I've Scaled

Infrastructure Scale

  • Auto-scaling web farms handling millions of requests
  • Batch processing systems crunching terabytes overnight
  • Multi-region deployments with sensible data locality
  • Database sharding when vertical scaling hit limits
  • CDN strategies that made global apps feel local
  • Monitoring systems that made the visible visible

Team Scale

  • Led distributed teams of developers across time zones
  • Built dev processes that survived 3x team growth
  • Mentored junior devs into seniors who could run things
  • Established technical governance without becoming bureaucracy
  • Created knowledge-sharing systems that actually worked
  • Built remote-first cultures before remote was mandatory

AI Infrastructure Design

MCP Servers & Gateways

  • Designed MCP servers that expose codebase structure to Claude through structured tools
  • Built gateway patterns that route agent requests to appropriate data sources
  • Created resource definitions that map complex systems to queryable AI context
  • Implemented authentication and authorization for AI agent access to internal systems
  • Architected rate limiting and guardrails for production AI infrastructure
  • Designed caching strategies to reduce context retrieval latency

Multi-Agent Orchestrators

  • Built orchestrators that coordinate specialized agents for complex codebase tasks
  • Designed agent handoff protocols for multi-step refactoring operations
  • Created shared context patterns that let agents collaborate on large codebases
  • Implemented task decomposition patterns that break complex work into agent-sized chunks
  • Built verification agents that review and validate generated code changes
  • Designed fallback and recovery patterns for failed agent workflows

Tech I've Worked With

Where I Can Help Immediately

  • Frontend: Vue.js, React, Angular—modernizing legacy SPAs
  • Backend: Java/Spring, Node.js, Python services
  • Databases: PostgreSQL, MySQL, MongoDB migrations
  • Cloud: AWS (Certified Professional), Terraform, Docker
  • Infrastructure: Kubernetes, CI/CD pipelines, observability
  • Modernization work: TypeScript migrations, design systems, platform upgrades

The Point

I don't just know these technologies. I know how they fit together, and I document that so your AI tools can use it too.

Got a Codebase That Needs Help?

Whether it's a rescue mission, a playbook project, or making your codebase work well with AI tools, let's talk.

Get in Touch