About

I am an applied AI researcher and engineer with a background in physics. My work sits at the intersection of retrieval-augmented generation, agentic LLM systems, and rigorous evaluation methodology. I am interested in building systems that are not merely performant but genuinely reliable, with measurement frameworks that can distinguish signal from noise.

Before moving into AI, I studied physics at the University of Melbourne, where my research focused on computational materials science. My master's work involved spectroscopic characterisation and defect modelling in single crystal diamonds, combining experimental measurement with computational approaches to understand nitrogen defect structures.

The transition from physics to applied AI was a natural one. Both disciplines demand careful attention to uncertainty quantification, robust experimental design, and scepticism toward results that look too clean. I carry these habits into my current work: building retrieval pipelines, designing evaluation harnesses, and developing agentic workflows for production systems.

My current focus areas include RAG system architecture and evaluation, dense retrieval failure modes, LLM-based planning and tool use, and the broader challenge of making AI systems that work reliably at scale. I write about these topics both on this site and at Chamomile.ai.

Education

• MSc Physics

  University of Melbourne

• BSc (Hons) Physics

  University of Melbourne

• Diploma in Mathematical Sciences

  University of Melbourne

Research Interests

• Retrieval-augmented generation and evaluation
• Agentic LLM systems and structured planning
• Dense retrieval failure modes
• Uncertainty quantification in ML systems
• Computational materials science
• Stochastic processes and optimal control