Work with me

I work with technology leaders in financial services and other regulated industries on AI systems they can measure, maintain and audit.

Agentic coding tools have made it easy to produce working software quickly. They have not made it easier to produce software you can maintain, validate or justify under regulatory scrutiny. Under pressure not to fall behind, teams are unleashing LLMs across their technology estates and discovering this the hard way. There is code no-one fully understands, a lack of systematic evaluation of whether the system does what it should and a lack of audit trail demonstrating why it behaves in the way it does.

I started in institutional finance, spending a decade building electronic trading and regulatory reporting platforms at tier-one banks including UBS, Deutsche Bank and ANZ, spanning cash equities, derivatives and OTC markets.

I spent the next decade making frontier technology production-grade for that same world. I created Web3j, the JVM Ethereum library used in production at J.P. Morgan and Samsung, which still sees over 4 million monthly downloads. I chaired the Enterprise Ethereum Alliance’s technical specification working group and delivered engagements with Microsoft, Vodafone and R3 through my company Web3 Labs.

I now work hands-on with agentic engineering every day and write about it here.

Most of my time goes on building and writing. Alongside that, I take on a limited number of engagements.

There are four ways to work with me.

Advisory call

A 90-minute session where I can answer specific questions for you. Examples include your agentic workflows, evals, architecture for AI-assisted delivery, or a second opinion for key decisions. You come away with clear recommendations and an additional perspective, alongside a recording and summary you can circulate to your team.

Fixed fee, £600 inc. VAT [book here].

AI code and evals review

A fixed-scope review for teams whose AI initiatives are outpacing their quality controls.

AI creates risk in two places: in your developer toolchains and in your product. I assess how your team uses AI-assisted development and the state of the code it produces.

Where AI sits in the product itself, in agents, LLM-powered features or automated pipelines, I assess whether you can measure its behaviour, which is the question most teams have not started on.

You get a written assessment, a prioritised list of what to fix first, and a debrief call with your team. Where there is more to do, the review doubles as the scoping phase for a larger engagement.

Fixed fee, from £5,000.

Fractional advisory

An ongoing engagement for leaders who want independent technical judgement on retainer. Decisions on architecture, tooling, hiring and vendors can be de-risked before they get expensive. Capped number of days per month, agreed in advance.

From £4,500 per month.

Larger engagements

Some problems need more than a review. For instance, standing up an evals capability for your team, or leading the remediation of an AI-assisted codebase. These are scoped individually, with capped days and a defined outcome.

Other fees exclude VAT where applicable.

I also advise on digital asset and blockchain initiatives. I spent the prior decade in that world, working with some of the world’s best-known companies and creating widely used developer infrastructure.

To kick off a conversation, email me via hi@conorsvensson.com, or book a short intro call.