About

I'm Seiya Alger-Hilton, an engineering manager with twenty years across fintech, telecoms, and energy, currently working in FX trading systems.

The thing that has followed me across all of it is the same: the places where technical decisions have consequences nobody has fully accounted for. Not the loud failures, which get post-mortems and regulatory scrutiny. The quiet ones. The systems that ship, the decisions that get made, and the gap between what was claimed and what is actually happening that nobody closes, because nobody is quite responsible for closing it.

I've watched that gap appear in financial risk systems that satisfied their model validation requirements without satisfying reality. In cybersecurity programmes that achieved certification while remaining genuinely insecure. In regulatory responses that addressed the letter of a requirement while the spirit went unmet. The pattern is consistent across industries and decades, and it's not subtle once you've seen it a few times.

AI is the largest, fastest deployment of consequential technical decisions in a generation. That felt like somewhere worth paying attention.

Hilton Labs is my outlet for thinking carefully about what that means, specifically around AI assurance, governance, and safety. Not in the existential sense, but in the institutional one: how do organisations evaluate systems they didn't build? How do regulators assess claims they can't verify? What does responsible deployment actually require, and how would you know if it was happening?

It's not a frontier research lab. The name is a mild affectation, a way of giving the work a container that feels appropriately serious without overstating what it is. What it actually is: one person reading widely, forming views, and writing them down. With a long-term hope of turning that body of work into something more, whether consultancy, independent research, or something not yet named.

My lens is that of a practitioner. I've spent two decades building and managing systems where "trustworthy" has a legal meaning, not just a marketing one. I'm not an AI safety researcher. I'm someone who has watched serious safety requirements get treated as paperwork in industries that should have known better, and who suspects AI is next in that sequence.

In 2026 I completed the BlueDot Impact AI Safety course and attended AI Safety Camp, less as credentials than as a commitment to engaging seriously with the field before writing about it.

That suspicion, and that engagement, are what Hilton Labs is here to work through. Slowly, honestly, and without pretending the answers are simpler than they are.

Hilton Labs Ltd is registered in England and Wales.

Correspondence welcome, particularly from researchers, regulators, and practitioners in adjacent fields. seiya@hiltonlabs.org