Experiment 001
The RTX 5090 context reader
Models, quantization, engines, chunking, and measured end-to-end wall time. Read the method alongside the results and download the public measurements.
Jacob Jennings / engineering notes
Building cool stuff and sharing the occasional side quest here.
Too many side projects. AI, ML, iOS, embedded systems, developer tools, hardware, gaming, CAD/CAM, and CNC.
By day: Apple Frameworks team at Google.
Opinions are my own.
Read the report →