Agents
AI Agents Orchestration
Agents are software, not magic. I design orchestration layers where AI agents plan, call tools, hand off to each other and escalate to humans, with state, retries, permissions and tracing built in from the start.
Core
Multi-agentTool callingMCP
Also in the toolbox
Claude tool useOpenAI AgentsLangGraphMCP serversTemporalCeleryOpenTelemetry
What you get
Deliverables, not just hours.
01Agent architecture & task decomposition
02Tool / function interfaces to your APIs
03MCP servers for your internal systems
04Human-in-the-loop approval flows
05State, memory and retry strategies
06Tracing, cost and quality monitoring
Typical use cases
Where this creates value.
Ops copilots
Agents that triage tickets, update CRMs and draft replies for human approval.
Research agents
Multi-step agents that gather, verify and compile reports from many sources.
Workflow agents
Agents that coordinate actions across Slack, email, ERP and internal tools.
Explore other services