Decades of hard-won expertise,
making the intractable,
testable.

Novamine pairs an adversarial multi-agent engine with decades of hard-won expertise in science, finance, risk, and delivery, turning the problems that stall serious R&D into verifiable, testable programmes of work.

Meet the team Talk to us

> Novamine, in numbers

400+ specialised agents in the engine
£15M+ global programmes delivered
100+ years of combined leadership
20 yrs at UBS Investment Bank

> The minds behind the method

Rob Panton

Rob Panton

Founder & CEO

Rob brings 25 years across hardware, network, and software engineering, and global IT programme delivery on projects over £15m spanning the UK, Europe, and North America. Before Novamine, he built an AI-driven therapeutic conversation platform in the pre-GPT era, solving ethical compliance, human-like dialogue management, and conversation memory before tools like LangChain existed. He designed and built Novamine's Structured Invention Engine from first principles: more than 400 specialised agents in a proprietary, multi-stage pipeline, investing £50k of personal savings and a year's unpaid leave to develop it.

LinkedIn
Piero Grieco

Piero Grieco

CFO / COO

A start-up CFO and advisor with a financial-markets background. Piero spent 20 years at UBS Investment Bank and qualified as an ACA with Deloitte, bringing financial rigour and operational discipline to every engagement.

LinkedIn
Gareth Thomas

Gareth Thomas

Commercial Lead for Scientific Innovation

An innovation and R&D leader who turns scientific ideas into commercial outcomes. Gareth's background spans GE, Alstom Power, Innovate UK Catapults, and the West Midlands Innovation Alliance.

LinkedIn
Dr. Christian S. Pedersen

Dr. Christian S. Pedersen

Commercial Director / Lead for Advisory

A strategy, technology, and transformation leader and advisor. Christian's career spans Oliver Wyman (20 years) and Accenture (5 years), alongside academic roles at the LSE, Cambridge University, and Imperial College. At Novamine he leads advisory relationships and commercial strategy.

LinkedIn

> We built the engine we couldn't find

Large language models are extraordinary, and structurally conservative. Trained to please and aligned to precedent, they gravitate towards the familiar, collapsing towards safe, average, recombinant answers. For genuinely hard R&D, that instinct is precisely wrong.

Novamine was founded on a different premise: that modern AI can be engineered to do the opposite. To leave the well-trodden regions of the data graph, generate radically original hypotheses, then force each one to survive adversarial scrutiny.

The output isn't persuasive prose. It is a verifiable, falsifiable programme of work, complete with controls, timing, and pass/fail criteria that a lab or engineering team can actually run.

That conviction is only half the story. Novelty becomes real when it meets human judgement: expertise, experimental reality, and commercial sense. Which is why Novamine is built by people who have spent careers turning complexity into outcomes, and why we work human-in-the-loop, from concept to lab.

“

Breakthrough-grade problem-solving can be engineered on demand, and proven with fast, falsifiable wins.
Rob Panton

Founder & CEO

> Principles we don't bend

01

Original by design

We reject safe precedent and prior art. Every engagement pushes into the uncharted regions of the problem space, where the most valuable answers are rarely the obvious ones.

02

Verifiable, not vibes

Persuasive prose isn't a result. We deliver falsifiable programmes of work: mechanisms, controls, and pass/fail criteria your team can put straight on the bench.

03

Your IP, protected

Every engagement runs under strict NDA. Your data is never retained or reused, and generated IP is transferred exclusively to you, or licensed on terms agreed in advance.

> Let's discuss your most intractable challenge.

Talk to us