Find where AI actually helps
Identify workflows where agents can meaningfully reduce manual work. Evaluate feasibility, expected benefits, implementation complexity, operating cost, and where agents do not make sense.
Work With Me
I build infrastructure for AI agents, with a focus on making long-running agent work persistent, observable, and recoverable.
What I help with
Design the right approach
Design an AI agent strategy around existing business and engineering workflows, including system design and human-agent collaboration.
Build and integrate
Build and deploy production agents connected to existing tools, systems, data, and processes.
Make adoption work
Train teams to work effectively with agents, avoid unreliable automation and unnecessary complexity, and measure whether the system saves time, reduces cost, or improves outcomes.
Engagements / Pricing
Architecture / Technical Review
I review one agent workflow or system across architecture, failure modes, observability, recovery, human handoffs, and operating cost. You receive written findings, prioritized fixes, and a 60 minute walkthrough. Typical turnaround: 5 business days.
$3,000
Implementation Engagement
For teams that need hands-on engineering after the review, I design and implement agent infrastructure inside the existing stack. Scope may include orchestration, persistence, recovery, observability, tool integration, operator workflows, and deployment. Defined after a technical fit call.
Custom
Proof
Agent Orchestrator
I have worked across Agent Orchestrator's backend and runtime systems, including session lifecycle, recovery, persistence, packaging, and reliability. I now focus on product and UX for supervising long-running agents without losing context, control, or trust.
Technical explanations / videos
I occasionally publish technical explanations of systems I build, design decisions, and ideas around AI agents and infrastructure.
AgentLab
AgentLab is the umbrella for the projects I build for the AI agents I use.
Who This Is For
- Engineering-heavy teams in any industry that are ready to use AI agents for real operational work, not demos or hype.
- Engineering teams running long-lived agents in real workflows, with an engineering owner responsible for the system.
How Engagements Work
- 01Understand the technical problem
- 02Inspect the existing system
- 03Identify architecture / reliability issues
- 04Recommend or implement changes
- 05Leave behind maintainable engineering work