Founder, AsymmetrIQ Labs
I run a quantitative research laboratory working at the intersection of machine learning and financial markets. The question underneath the work is a general one: can a system learn to act sensibly in an environment that is adversarial, non-stationary, and unforgiving of overconfidence?
I came to it from institutional analysis rather than from engineering, and that shows in what I find interesting about it. The part of an autonomous system I care most about is not what it can do. It is what it is not permitted to do, and whether that limit holds when the system is confident.
A quantitative research laboratory developing autonomous decision systems. The current system is in development. Its architecture separates the components that generate decisions from an independent governance layer that constrains them, and gives that layer real authority rather than an advisory role, on the principle that a constraint which loses to a sufficiently confident model is not a constraint. asymmetriq.ai
Contributor to an open-source autonomous-agent project. My merged work has centred on the agent discovery layer: the on-chain registry lookups that let independent agents find one another. View the project
More at asymmetriq.ai.
Work that sat where regulation meets operational reality: multi-jurisdiction payroll, equity administration, and the controls that keep both auditable under scrutiny. It is a useful place to learn that a control nobody can enforce is documentation rather than a control, which is the same problem autonomous systems have.
Institutional analysis at bottom: how rules, incentives, and enforcement produce outcomes in complex systems full of participants pursuing their own interests, and why formal authority so often diverges from effective authority. Markets pose the same problem in different dress.
Quantitative research. Machine learning. Autonomous systems. Systems governance. Financial markets. People operations. Labor studies and employment relations.
LinkedIn is the most reliable way to reach me.