AgentRepoCoach
Open source · PyPIAn open-source tool that scores codebase health for AI-agent readiness, published on PyPI and GitHub for anyone to inspect.
Cloud-platform and IT4IT product leadership inside large Dutch enterprises, in health insurance and transport. That record sits on an engineering foundation from banking and government: regulated environments, stakeholders who rarely all agree, roadmaps that have to be funded and defended. That is the work I do as a senior product owner and product manager.
Real hands-on depth in AI and cloud systems sits under that product work. On my own account I design and build multi-agent AI frameworks and LLM orchestration. What an AI initiative costs and whether it can ship are product decisions before they are engineering ones.

The Azure cloud platform I own today sits inside a large Dutch health insurer’s Cloud Center of Excellence. Roughly thirty internal teams build and deploy on it, and as product owner I set its direction and lead the ten engineers who build it.
Before this came product ownership at a national transport operator, where I led the automation team and, when the programme needed senior leadership, stepped up to run additional teams and the general direction of its IT4IT platform. The engineering underneath the product work came first: developer roles in banking and government, inside the same regulated walls. Compliance and audit have shaped the work firsthand ever since.
Leading means getting specific: reading the system, weighing the risk, putting the trade-off in front of stakeholders in plain terms. When a delivery decision needs challenging, I get into the detail with the engineers, then turn it into a plan the business can fund.
BrainSpark Innovations is deliberately a one-person practice: senior product leadership on every engagement, and fewer of them at once so each gets full attention. The product decisions stay close to the detail.
My hands-on work is my own track record. Three that show the range are below.
An open-source tool that scores codebase health for AI-agent readiness, published on PyPI and GitHub for anyone to inspect.
A production multi-agent AI orchestration framework, released continuously on the Anthropic Claude and OpenAI APIs, reaching from agent decision loops down to the cloud-native platform patterns and infrastructure-as-code beneath them. Built and run on my own account.
A multi-agent autonomous trading platform in C#/.NET, with a specialised agent fleet and a full automated test suite. Built, tested and deployed end to end, to the standard an audited environment demands.
Tell me what you are building and where it stands. You will get a straight product read back.
Currently engaged on a senior product-ownership programme. Open to conversations about what comes next.
References available on request.