Engineering notebook
Samuel Diop
Machine Learning Engineer | GPU Optimization & Efficient AI | PhD
I build and optimize machine learning systems, with a focus on GPU computing, efficient inference, and computer vision.
This website documents my engineering work, technical experiments, and research, including the decisions, trade-offs, and lessons behind them.
Selected Work
Can ordinary video detect infantile epileptic spasms?
Learned action recognition and classical optical flow evaluated on long clinical videos.
Read engineering record →Can Echo State Networks become practical at the edge?
A C++17 inference engine with custom CUDA kernels and sparse reservoir storage.
Read engineering record →Can action recognition run entirely in a browser?
X3D-M through WebAssembly, without a backend or video upload.
Read engineering record →Latest Writing
Why asynchronous model loading was three times slower
Negative result
How to report temporal-model throughput without ambiguous FPS
Benchmark
What a 10× pipeline speedup does and does not prove
Performance
View all writing →