Building runtime containment for autonomous AI agents. Budget limits, circuit breakers, trust boundaries -- the layer between your agents and your API bill.
veronica-core -- Runtime containment for LLM agents (v3.0)
- Hard budget enforcement (HALT, not alert)
- Step limits, loop control, circuit breakers
- A2A trust boundaries, policy simulation, async MCP adapter
- AG2 / LangChain / CrewAI integration
- 3900+ tests, 92% coverage
pip install veronica-core
Construction infrastructure consultant turned AI engineer. 10+ years in regulated, zero-tolerance environments where failure has real-world cost. Now applying that mindset to LLM agent safety.
Python, C++, CUDA, TypeScript, WebGPU
Runtime enforcement, multi-agent orchestration, GPU systems, 3D graphics pipelines
Available for remote contract work -- AI safety, agent infrastructure, runtime systems.