AUTONOMOUS-FOX
LABORATORIES
About us
What we do....
We’re Autonomous-Fox Laboratories, an independent deep learning research lab focused on advancing AI for financial market modelling. Founded in 2025, the company is in its early stages operationally, but its research foundation is not. Our work builds on more than 12 years of applying machine learning to financial markets, spanning systematic trading, portfolio construction, statistical arbitrage, and market making.
Financial markets are without doubt one of the most hostile environments for data science. They challenge all standard assumptions because the statistical properties of the data change continually and the signal-to-noise ratio is exceptionally low. Consequently, we must avoid generic solutions. At AF Labs, we design highly specialised architectures to disentangle signal from noise and infer the hidden structures driving asset prices.
Some are bulls, others bears.
We are foxes, we craft intelligent systems to discover opportunities in all markets.
Shaping the next generation of trading systems with our foundational models of market behaviour
Embracing Complexity
Markets are not just a collection of independent prices; they are adaptive webs of interaction. Our models capture interdependencies and transient relationships, uncovering mispricings that emerge when complex equilibria are temporarily disrupted.
Going Deeper Than Ever Thought Possible
Over the past decade, deep learning has revolutionised entire industries — while finance sat on the sidelines watching and waiting. AI has transformed vision, language, and biology, however financial modelling was left behind. The reason is simple: markets pose a much harder problem. Data is messy, relationships unstable, and structure always changing. Deep learning wasn’t built for that world — not then. But now that’s changing. Advances in machine learning and neural methods more generally, are finally giving us tools that can understand markets on their own terms, allowing us to build systems that can learn meaning, not just basic patterns. In 2025, the gap between deep learning capability and market reality is closing fast, and a new generation of quantitative trading models is about to emerge. At AF Labs, we intend to be right at the front of that renaissance.
Behaviours Over Statistics
Our approach is centered on behavioral modeling. We view markets as a collection of assets exhibiting quantifiable behaviours. Traditional statistics cling to assumptions of stability and order that markets rarely honour. They describe what has already happened and infer that it will happen again. Strip away the assumptions, and you’re left with noise — not foresight. Instead, we believe that behaviors constitute the very language of the market, and how those behaviours play off one another over time is the story. Behavioural modelling represents the next paradigm for understanding complex systems — and deep learning is the mechanism through which we can learn to interpret that language.
A Model for All Markets
We remain cautious using the F-word --we're not quite there yet, but our design philosophy of behavioural modelling is deliberately agnostic to any single asset or market.
Rather than fitting to surface patterns, our models seek the latent structures that govern how systems evolve and adapt — a framework capable, in principle, of generalising across the entire market ecology.
In this view, behaviours form the vocabulary of the market's language, and their interactions become the conversations through which structure emerges, which can then be followed by comprehension.
Dare we suggest, we are on the way toward a foundation model for financial markets.
Experience and Expertise
16+
Years of Experience in quantitative financial market modelling
12+
Years of Experience in Machine Learning
100s
Trading Models Developed
100%
of researchers are PhDs
1
Objective
Join the Team
We are looking for exceptional machine learning experts to join our small but growing team of research scientists at a founding level. In particular, we seek specialists in representation learning and probabilistic / generative modelling — in other words, wizards with the ability to bend neural architectures to their will.
If you hold a PhD in computer science or a related field (theorists included) and want to tackle some of the hardest problems humanity has ever faced, we’d love to hear from you.