Bijan Bowen is an AI consultant and technical educator in the Boston area. His tutorials and videos examine how models behave in practical tasks and document the setup required to run AI tools on accessible hardware.[2][6]
Start with BitNet b1.58 on Raspberry Pi 4B →
Links
Owned technical writing Owned professional biography
Owned linked GitHub account 
X · bijanbowen Career
Bowen describes consulting with small businesses on integrating AI into their workflows and runs the educational channel bearing his name. His public material combines configuration guides with hands-on coding and creative tasks, giving viewers a practical way to investigate a tool beyond its announcement or benchmark summary.[2][6]
Contributions
Make an ARM inference build reproducible
His BitNet setup guide documents a tested Raspberry Pi 4B configuration using 64-bit Raspberry Pi OS and two gigabytes of RAM. It installs compiler and Python dependencies, generates the tiled lookup-table headers required by the low-bit kernels, builds the runtime and downloads the model. An interactive inference command then checks that the model, library and Python binding work together. The explicit code-generation step addresses a missing-header build failure. The April 2025 guide describes its tested software environment.[5]
Compare models on recognizable application work
His GLM 5.3 Flash video’s published test plan spans browser workflows, generated games, C++ implementations, model training and a Blender/Godot example. Those named tasks make the evaluation setup easier to inspect than a single general rating. The video presents his hands-on comparison with the full model; its original description identifies the tasks and the result-review segment.[6]
Tools, writing & talks
Writing, research & talks
Start here
BitNet b1.58 on Raspberry Pi 4Bbijanbowen.comA step-by-step guide to compiler setup, low-bit kernel generation, model download and an interactive inference test.
Read the tested hardware and OS prerequisites, then follow the kernel-generation step before building the runtime. The guide assumes Linux terminal and Python-environment familiarity; its exact commands concern the documented April 2025 toolchain.
Author of setup tutorial[5]A hands-on model comparison using coding, agentic and creative tasks identified in the original video description.
Presenter and channel creator[6][2]Original source