The Journey: Analytical Rigour to Autonomous Intelligence
I started in mechanical engineering, earning a First-Class Honours degree with an emphasis on mathematical modeling, thermodynamics, and fluid mechanics. Engineering taught me the fundamentals of physical constraints, systematic diagnostics, and numerical optimization—principles that directly shaped my transition into high-performance software engineering and machine learning.
At V-Nova, I operated at the intersection of video compression codecs, GPU pipelines, and distributed cloud computing. Collaborating with Meta and NVIDIA engineers, I designed PaPeR, the automated testing harness that reduced distributed experimentation cycles from three hours to seconds and enabled scalable VMAF/PSNR benchmarks across thousands of video streams.
At Roku, my focus evolved to the cutting edge of AI-assisted software delivery: building Roku Explorer (an autonomous exploratory QA platform using reinforcement learning, multimodal vision, and ChromaDB RAG to explore device UIs), QAntum (a Slack-native multi-agent orchestrator connecting specialized agents via capability protocols), and Forge (an AI agent authoring regression test suites in isolated git worktrees). This work represents a patent-pending innovation (US/EU) for autonomous software verification.