Kristen HauptFounder, Violet SystemsI have spent my career in places where a bad decision doesn't simply mean a bad quarter or year. It can mean a patient receiving the wrong medication, a credit model making unfair decisions that affect real families, or a member whose data is not protected. Working in the most regulated industries has taught me that technology is only as good as the decisions it makes possible, and only as valuable as the trust people can place in those decisions.
Along the way, I kept seeing the same gap, and the same opportunity. The people who build systems, the people who manage risk, and the people accountable for results rarely sit at the same table. Engineers want to move fast and build the latest and greatest tools. Risk and compliance teams slow things down to protect the organization. Business leaders want to see returns. All three are right, and AI has made that gap wider and the stakes higher.
I started Violet Systems to close the gap. As someone who sits at all three tables, I write the code, design the data infrastructure, and build the models. I also sit down with executives, auditors, and regulators to explain why a system does what it does, and why metrics can look bright green while the business is still bleeding. Governance is not a document handed over at the end of a project. It is built into the system every step of the way.
As far as the name, I just really like purple.
Before choosing a tool, I want to understand what you are trying to decide, what evidence you have, what could go wrong, and why it matters.
If you cannot explain how a system reached its answer, you cannot defend it, improve it, or fully trust it.
Done well, guardrails let teams ship with confidence instead of waiting on approvals and rework.
I build the smallest thing that answers the question well, prove its value, and grow it from there.
The most expensive tool is not always the best fit. I start with the objective, gather the surrounding details and weigh the options before recommending one.
AI agents are starting to take actions, not just make suggestions. My research focuses on reusable patterns that let those agents work inside real-world rules: clear permissions, human checkpoints and a full record of every decision. The ideas I test there find their way into what I build for clients.
I would love to hear about it.