The engine behind the mission
To serve thousands of athletes at scale, we have to be extremely efficient – and that starts with technology built for purpose. ARC’s tech isn’t just a tool; it’s a mission-critical engine designed to support the often-ignored 95% of athletes who leave competition without a roadmap.
At the heart of this is our custom AI – trained on a private LLM stack built on NVIDIA architecture. This isn't just a chatbot. It’s a digital mentor, strategist, and creative partner, trained to help athletes define their brand, grow their audience, and collaborate with each other to build a post-sport career with purpose.
From content creation to profile optimisation, learning pathways to business opportunities, ARC’s AI takes the heavy lifting out of reinvention – and puts actionable insight straight into the hands of athletes. It’s built to listen, learn, and grow with our community – becoming smarter and more valuable with every interaction.
Our technology
ARC is being built on a modern, modular architecture that balances scale, speed, and personalisation. Our front-end is React-based with a component-driven design system, enabling rapid iteration and seamless integration of new features. The core platform is engineered with automated and AI-assisted development workflows, ensuring velocity without sacrificing code quality.
The backbone of our intelligence layer is a proprietary orchestra of SLMs, fine-tuned in a containerised training environment using NVIDIA GPUs. We are not relying on third-party endpoints – we’re building our own models, tailored specifically for athlete use cases. We are using small, generalist base models, then create the lion share of the final model using our highly curated, specialised domain data.
Using a combination of supervised fine-tuning (SFT), retrieval-augmented generation (RAG), and prompt chaining, and layered on state of the art architecture - that develops as the technology does - our models are optimised to deliver high-context, high-relevance outputs across brand strategy, content creation, and learning support.
Our architecture supports agentic workflows – allowing athletes to delegate complex tasks (like brand audits, social planning, assisting in creating business plans, or simply to give advice on generalist life skills) to autonomous agents that can reason, iterate, and refine output over time. These agents will eventually act as personalised operating systems for each athlete’s post-sport journey – learning from behaviour, adapting to goals, and enabling collaboration within the wider ARC ecosystem.
We’re also investing heavily in back-end extensibility, with a cloud-native stack built for global scale and granular permissions – ensuring privacy, compliance, and future integrations with third-party services like financial advisors, brands, and educational platforms.
This isn’t off-the-shelf tech. It’s purpose-built infrastructure designed to transform under-served athletes into high-performance operators – and to scale that impact across thousands of users without compromising depth, nuance, or control.