Sybil Solutions

Public Track Record

Work Completed

Everything built, released, taught, and given away. No marketing fluff, just the artifacts.

Hardware Donated

GPUs

2x RTX 3090 (48GB VRAM)

One recipient active on X

Laptop

1 Student Laptop

Student without hardware access

Education

Algerian Students

4 Classes

Weekly sessions

X Followers

2 Classes

Discord community sessions

Total Students

6 Cohorts

Completed

Model Compressions & Releases

Compressions Released

18

Tested & working

Total Downloads

100k+

Across all releases

vLLM TurboQuant

First Implementation

1,400 GitHub stars

Open Source Contributions

vLLM Studio

50+ Contributions

Free & open source, actively used

pi-brain

Session Extraction Tool

Extract and share AI sessions

reap-mlx

MLX Compression

Compression on Apple Silicon

reap-strix

Strix Compression

Compression on any hardware

Research Support

Student Researchers

2 Supported

1000+ compute credits

Prime Intellect

$100 Personal Grant

Follower compute access

Published Wikis & Documentation

RL Wiki

Basic RL Concepts

rl-sybil.vercel.app

Strix Benchmarks

Inference Optimizations

strix-benchmarks.vercel.app

Blackwell GPU Wiki

SM120 / CUDA Pipeline

blackwell-gpu-wiki

Technical Milestones

Day 0 Support

DeepSeek-V4-Flash

SGLang on SM120 (Blackwell)

Talks Given

Multiple Events

Public presentations

Droid Guide

Free Usage Article

How to use Droid for free

Verification & References

Etherscan

Donation Transaction

View on Etherscan

X/Twitter

Status Reference

View Tweet

vLLM Studio

GitHub Contributions

github.com/0xSero/vllm-studio

Profiles

Shared 100+ pi-sessions publicly. Working daily on local AI sovereignty: freedom over dependency, durable capability over recurring rent.