01. identity

artificial intelligence architect

i build robust pipelines, train sota models, and open-source the future. my work bridges the gap between complex ml theory and usable, functional applications.

view operations github

console.log(status)

const status = {
  active: true,
  training: ["i3-1B-Ethan", "i3-Pocket"],
  uptime: "99.9%",
  location: "remote"
};

// current task
optimize(i3_architecture);

02. current_state

active_training_sequence

gpu_utilization 84%

03. development_index

selected projects

project_01

i3-lab

state-of-the-art pre-training language model pipeline. efficient, scalable, and built for experimentation.

[collection] [hf space]

project_02

open-i3

the open-source release of the i3 architecture. promoting transparency in ai development.

[github repo]

project_03

ChessSLM

small language models trained exclusively to play chess. available variants: base, rl, neo, and nano.

> play_against_them

project_04

i3-papers

documentation & official research papers. technical specifications and architectural overviews.

project_05

literallynothing.js

npm package that does exactly what it says. 0 bytes of utility. functional minimalism at its absolute limit.

04. model_zoo

hugging face models

model_id	params	ctx_window	hardware	access
i3-4096ctx	~300m	4096	nvidia p100	view_card
i3-200m v2.0	~200m	2048	v100	view_card
i3-80m	~80m	1024	legacy	view_card
i3-22m	~22m	512	legacy	view_card

more to be presented soon...

05. capabilities

technical capability stack

01. language python 3.10+

02. framework pytorch / cuda

03. distribution huggingface

04. processing hf datasets

06. contributions

fuel the research

computing power is the primary constraint. if my models aided your development, consider supporting the next training sequence.

support_via_kofi vast.ai_template