Senior AI Architect & Evangelist @ Solo.io
Building the infrastructure and connectivity layer for the agentic era.
I'm a technology leader, architect, developer advocate, and open source contributor focused on the infrastructure behind the agentic AI era.
Today, I'm a Senior AI Architect & Evangelist at Solo.io, exploring how we build, connect, secure, observe, and govern AI agents in production.
My work sits at the intersection of:
- π€ Agentic AI & AI infrastructure
- π APIs, MCP & agent connectivity
- π Cloud-native & distributed systems
- 𧩠Agent platforms, gateways & governance
- β‘ Event-driven architectures
- π οΈ Open source & developer ecosystems
Agents are not just a model problem. They're an infrastructure problem.
The industry spent decades building the connective tissue of modern software: APIs, identity, gateways, service meshes, event streams, observability, security, and cloud-native platforms.
Now agents are becoming a new kind of consumer of that infrastructure.
The opportunity isn't to throw away everything we built. It's to evolve it for a world where software can reason, discover capabilities, and take action on behalf of people.
I think the agentic stack increasingly looks something like this:
βββββββββββββββββββββββ
β Agents β
ββββββββββββ¬βββββββββββ
β
MCP Β· A2A Β· Agent APIs
β
ββββββββββββΌβββββββββββ
β AI Connectivity β
β Gateways Β· Routing β
β Identity Β· Policy β
ββββββββββββ¬βββββββββββ
β
APIs Β· Events Β· Tools
β
ββββββββββββΌβββββββββββ
β Cloud Native β
β Services Β· Data β
β Infrastructure β
βββββββββββββββββββββββ
The interesting problems are therefore not only about models.
They're about connectivity, context, identity, security, governance, observability, reliability, and cost.
And that's where my work is focused.
- How should APIs evolve when their primary consumers are agents?
- Where does MCP fit in the enterprise architecture?
- How do agents discover and securely access capabilities?
- How do we govern agent-to-agent and agent-to-tool interactions?
- How do we manage context without creating context bloat?
- How do we control the cost and complexity of agentic workloads?
- How can event-driven architectures make agents more reliable and autonomous?
- What can we learn from decades of API management and cloud-native infrastructure?
I've spent much of my career working with and contributing to open source and developer communities.
Areas I've contributed to and explored include:
- π¬ Microcks, API and microservices mocking and testing
- π AsyncAPI, event-driven architecture and standards
- βοΈ Cloud-native developer tooling
- π€ AI, MCP and agentic infrastructure
- π§ͺ API testing, contract testing and developer workflows
I'm especially interested in projects that make complex infrastructure easier for developers to understand, adopt, and operate.
I'm building The Agentic Architect, where I share practical ideas about:
AI + APIs + Agents + Infrastructure
Architecture breakdowns, demos, experiments, MCP deep dives, conference talks, and observations from the road.
The agentic era needs more than intelligent models. It needs an architecture.
I regularly speak, write, teach, and participate in developer and cloud-native communities around the world.
Topics I enjoy exploring:
Agentic AI Β· MCP Β· AI Gateways Β· APIs Β· Cloud Native Β· Event-Driven Architecture Β· Open Source Β· Developer Experience
- π The Agentic Architect
- πΌ LinkedIn
- π GitHub
Building the connective tissue for the agentic era.