Imaging pipelines · Video encoding · GPU programming · Performance optimization · C++
Current stack
Experience
Research engineer exploring next-gen video pipelines, encoding, GPU & graphics (Vulkan/OpenGL), game engines, and AI/computer-vision pipelines — with performance at the centre.
Built ClickShare apps in C++/C/Kotlin with TDD — peripheral management, video streaming and camera capture on GStreamer. Worked in Scrum, occasionally standing in as scrum master.
Researched, prototyped and implemented the noise-reduction algorithm for a speed-focused RAW editor in C/C++ with Halide, tuning quality and performance to the underlying CPU/GPU.
Skills
Languages & tools — in order of proficiency →
Spoken
Projects
Optimised a light-field renderer (the technology enabling 6 degrees of freedom for camera-captured content), written in C/C++. Profiled CUDA kernels to find bottlenecks and fix them — maximising warp occupancy and optimising memory accesses.
A Rust solver for bin packing with multiple possible bin sizes, paired with a React front-end that talks to it over gRPC / REST.
General-purpose GPU image histogram equalisation written in C++ with OpenCL.
Implemented parts of an H.264 video codec: transform, quantisation, intra-prediction, and motion estimation/compensation with partitioning.
Implemented and statistically compared kNN, random-forest and neural-net classifiers in TensorFlow to predict cell type from single-cell RNA data.
A Python/OpenCV tool that reconstructs a visitor's walking path through a museum from video footage, aligned against a known floor map.
Contact