sekka

Python, Java & Spring (Boot, Spring AI), JVM ecosystem. LLM integration, multi-agent orchestration, agent governance & LLM security. More specialist than generalist — depth on the AI-runtime side.

The unglamorous runtime side of multi-agent LLM systems — durable execution, reason-aware retries, agent governance — plus the security surface of agentic AI (prompt injection, tool misuse).

AgentFlow4J — an open-source orchestration runtime for governed, stateful multi-agent LLM workflows on Spring AI: graph execution, durable runs, and budget/tool/approval gates.

Open-source collaboration on AgentFlow4J, and chats about multi-agent orchestration, agent governance, LLM security, or building AI agents on the JVM. Say hi if you do agents in Java/Spring!