Hi, I'm James

I build reliable machine learning systems using high-performance computing principles on small, fully reproducible projects. My method is clear: define problems tightly, document assumptions, and benchmark against simple baselines before deeper optimisation. I address real-world constraints up front—data quality, latency, throughput, reliability, and cost. My workflows stay straightforward, with clear notebooks, version-controlled environments, and concise logs. When issues arise, I write brief postmortems to guide the next iteration.

I am exploring how to make AI systems cohesive, connecting data pipelines, training loops, deployment, and monitoring. My priorities are practical workflows, APIs, batch job management, and automated scheduling. Right now I am sharpening core technical skills, focusing on Python, numerical computing, statistics, linear algebra, and benchmarking for small projects.

After my BSc, I aim to pursue a PhD in ML systems or AI architecture. My long-term goal is to design AI systems that are dependable, efficient, and cost-effective in real-world environments.