What are the responsibilities and job description for the AI Researcher position at Smart Bricks?
About Smart Bricks
Smart Bricks is a frontier AI lab building autonomous reasoning systems that allow capital to discover, evaluate, and transact assets end-to-end. We sit at the intersection of frontier AI research and one of the world's largest and most data-rich industries - and we are building infrastructure that compounds in capability with every decision the system processes. Research here is not academic - it goes directly into production and has real consequences at scale.
About The Role
As an AI Researcher at Smart Bricks, you will advance the research agenda of our model and intelligence layer - discovering approaches that work at scale and taking them from idea to production. We are looking for people who want to work on problems that are both technically interesting and commercially consequential. You will own a research agenda, collaborate closely with engineering to take your work to production, and contribute to a broader vision of what intelligent autonomous systems can do.
What You Will Work On
Smart Bricks is a frontier AI lab building autonomous reasoning systems that allow capital to discover, evaluate, and transact assets end-to-end. We sit at the intersection of frontier AI research and one of the world's largest and most data-rich industries - and we are building infrastructure that compounds in capability with every decision the system processes. Research here is not academic - it goes directly into production and has real consequences at scale.
About The Role
As an AI Researcher at Smart Bricks, you will advance the research agenda of our model and intelligence layer - discovering approaches that work at scale and taking them from idea to production. We are looking for people who want to work on problems that are both technically interesting and commercially consequential. You will own a research agenda, collaborate closely with engineering to take your work to production, and contribute to a broader vision of what intelligent autonomous systems can do.
What You Will Work On
- Researching and developing new model architectures for prediction, classification, and decision-support tasks operating on proprietary structured, unstructured, and visual data
- Exploring and advancing graph neural network architectures for large-scale relational knowledge graphs - modelling entities, relationships, and temporal dynamics at scale
- Advancing our LLM layer - investigating fine-tuning approaches, RLHF methodologies, and retrieval-augmented generation for domain-specific reasoning and explainability
- Researching cross-domain model transfer - understanding how models trained in one context generalise to another, and designing architectures that improve cold-start performance
- Exploring multimodal learning - integrating image, text, and structured data signals into unified intelligence representations
- Publishing and presenting findings internally to drive model improvement and shape research direction across the team
- Have a track record of coming up with new ideas or improving upon existing ideas in machine learning - demonstrated through publications, projects, or production implementations
- Be able to own and pursue a research agenda - choosing impactful problems and driving them to completion autonomously over long-running projects
- Have strong fundamentals in deep learning with comfort in PyTorch, transformers, and graph neural network frameworks
- Be excited about research that ships - not just research that publishes
- Communicate research clearly to audiences with different backgrounds - from fellow researchers to engineers to non-technical stakeholders
- PhD in Computer Science, Mathematics, Statistics, or a related field
- First-authored publications at peer-reviewed ML conferences or journals such as NeurIPS, ICML, ICLR, or ACL
- Experience with graph neural networks, geospatial ML, or multimodal learning
- Background in applied AI research in a production or commercial setting
- Experience with reinforcement learning from human feedback (RLHF) or preference optimisation
- Solve Hard Problems:Work on AI systems that are live in production - agentic orchestration, real-time inference, cross-market model transfer, and retrieval systems operating at scale on proprietary data that doesn't exist anywhere else.
- Build What's Next:The infrastructure we are building sits at the frontier of applied AI. The models, agents, and reasoning systems you work on here will define how one of the world's largest asset classes operates for decades.
- Ownership and Impact:Small team, no bureaucracy, high trust. Your work ships, your decisions matter, and your fingerprints are on everything we build.
- Learn from the Best:Collaborate with world-class engineers, researchers, and operators who left careers at leading AI labs and financial institutions to build something genuinely new.