AI Expert – Trading Systems & Agent Development

Most AI systems never face reality. Ours do. We deploy into live trading environments where information is incomplete, conditions change quickly, and performance is measurable in real time. We are looking for someone to build systems that improve how decisions are made and executed under pressure — systems trusted because they work, not because they sound impressive. This role carries real ownership, direct commercial impact, and unusually fast feedback loops. The Opportunity You will work directly with traders and decision-makers to build AI systems that: - identify actionable signals faster - improve decision quality under uncertainty - compress the time between insight and execution - increase consistency and execution precision - continuously improve from live outcomes You will own systems end-to-end: problem framing → architecture → deployment → iteration The expectation is production impact. What Makes This Role Different Most AI roles sit far from commercial reality. This one does not. Your work will be deployed quickly, tested under live conditions, and measured against outcomes that matter. You will operate with: - direct access to decision-makers - minimal bureaucracy - high autonomy - high accountability - tight feedback loops - meaningful commercial exposure This is an environment for builders who want to see their systems used, trusted, and scaled. What We’re Looking For You have already built systems that operate reliably in production environments. You are comfortable working with: - incomplete information - imperfect data - unclear problem definitions - fast iteration cycles - high-stakes decision environments You move quickly. You think commercially. You care about measurable outcomes. You are likely stronger at shipping than presenting. Technical Background You will probably have strong experience across several of: - Python and modern ML infrastructure - LLM systems and agent architectures - orchestration, retrieval, and evaluation systems - reinforcement learning or optimisation systems - event-driven or streaming systems - low-latency production environments - real-time inference infrastructure Experience in trading, betting, or probabilistic decision systems is useful but not required. We care more about judgement, speed, and the ability to build systems that survive contact with reality. What Success Looks Like Within the first few months, we would expect: - multiple systems deployed into live workflows - measurable improvements against baseline performance - tighter decision loops - increased execution consistency - systems trusted by operators in production Success in this role is measured by commercial impact. Environment - Small team - High trust - Fast execution - Direct communication - High standards - Strong ownership culture We hire carefully and prefer a small number of exceptional people over scaling headcount. Compensation Compensation is flexible and designed to attract exceptional individuals.