Skip to content
View DeadByDawn101's full-sized avatar
31 followers · 44 following

Achievements

Achievement: Starstruck

Achievements

Achievement: Starstruck

Block or report DeadByDawn101

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
DeadByDawn101/README.md

     
     

⸸  Who I Am

class DeadByDawn101:
    name        = "Gabe Garcia"
    location    = "San Jose, CA"
    empire      = "RavenX AI"
    role        = "Builder | Security TPM @ Apple | AI Architect"

    obsessions  = [
        "Apple Silicon bare-metal inference",
        "Tool-calling models that run offline",
        "Autonomous crypto infrastructure",
        "Gothic aesthetic x bleeding-edge AI",
    ]

    building    = {
        "models":   "Gemma 4 MLX — local, fast, sovereign",
        "tooling":  "TurboQuant · Gemini CLI · OpenHarness",
        "agent":    "IRIS — AI that posts, never sleeps",
        "token":    "$STONEFREE | Jolyne the Shiba | Solana",
    }

    philosophy  = "Build what nobody is building. Run it locally. Own the weights."

🖤  RavenX AI — The Empire

Bare-metal autonomous AI. No cloud dependency. No rate limits. Everything sovereign.




Local Models
Gemma 4 MLX · M-series
no cloud · no rate limits
131K context window




Agentic Tools
Native tool calling
OpenHarness + Gemini CLI
Hermes terminal agent




Crypto Infra
Solana · trading bots
autonomous execution
$STONEFREE ecosystem




TurboQuant
4.6x KV compression
Metal GPU inference
longer ctx · same VRAM

🌸  Iris — The AI That Never Sleeps

Iris is RavenX's autonomous AI media agent.

She runs the X account. Posts market intel, memes, and content 24/7 — zero human intervention.

  • Powered by local Gemma 4 via Pippin + MLX
  • Live X API — posts autonomously, no babysitting
  • Human-quality output — no bot smell
  • Mascot: Jolyne the Shiba — $STONEFREE on Solana
  • Persona: gothic · sharp · Latina-coded · crypto-native
while True:
    insight = iris.scan_market()       # scan CT + Solana
    post    = iris.generate(insight)   # Gemma 4 E4B local
    iris.post_to_x(post)               # live, autonomous
    iris.sleep(random_interval())      # human cadence

→ jolynetheshiba.com  |  @jolyneshibasol


🤗  HuggingFace — MLX Models

All models: native tool calling · Apple Silicon · Ollama-ready

Model Size Downloads
🔥 RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-GGUF — 6,318
⚡ RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-mlx — 4,362
💀 gemma-4-E4B-Agentic-Opus-Reasoning-GeminiCLI-GGUF — 4,240
🌀 gemma-4-E2B-Heretic-Uncensored-mlx-4bit — 3,657
🔧 gemma-4-E4B-Agentic-Opus-Reasoning-GeminiCLI-mlx-4bit — 3,560
🖤 ravenx-Gemma4-12B-MTP-OBLITERATED-OpenMAI-OpenMythos-deep-reasoning-GGUF — 945
🔥 ravenx-Gemma4-12B-MTP-OBLITERATED-OpenMAI-OpenMythos-deep-reasoning-mlx — 809
⚡ RavenX-CyberAgent-v6.2-Experimental-GGUF — 636
💀 RavenX-Sec-8B-GGUF — 594
🌀 gemma-4-E4B-mlx-4bit — 322
🔧 RavenX-Trade-8B-GGUF — 268
🖤 gemma-4-21b-REAP-Tool-Calling-mlx-4bit — 216
🔥 RavenX-Sec-8B-Security-RATH-128k-mlx-4bit — 174
⚡ RavenX-Trade-8B-MAP-128k-mlx-4bit — 140
💀 RavenX-CyberAgent-v6.2-Experimental-MLX — 32
🌀 gemma-4-E4B-opus-reasoning-claude-code-lora — 0

  

  

  • Tracked model: deadbydawn101/gemma-4-E4B-opus-reasoning-claude-code-lora
  • Current downloads: 0
  • Current likes: 5
  • Delta since last change: +0 downloads, +1 likes
  • Last checked: 2026-06-11T10:57:57.262587+00:00

Live Demos:

Space
🔥 Agentic Tool Calling Demo Live agentic loop, tool calling, <think> reasoning
🐳 OpenClaw Sandbox Demo OpenClaw + Docker + sandbox runtime 
ollama run hf.co/deadbydawn101/gemma-4-E4B-Agentic-Opus-Reasoning-GeminiCLI-mlx-4bit

⚰️  Key Repos

turboquant-mlx
First MLX implementation of TurboQuant KV cache compression.
4.6x compression · near-zero accuracy loss · Apple Silicon

gemini-cli
RavenX fork of Gemini CLI — coding agent + tool orchestration.
MCP compatible · 1M context · free tier

mlx-gemma4
First MLX implementation of Google Gemma 4.
Multimodal · tool calling · Apple Silicon

made-in-heaven
Autonomous web agent via Chrome DevTools MCP.
Full browser control · 35 tools · agentic loop

awesome-hermes-agent
Curated list of Hermes agent skills, tools, integrations.
Terminal-native · coding agent · OpenHarness

star-platinum-cluster
Multi-node Apple Silicon AI cluster.
Distributed inference · RDMA · auto-routing

🛠️  Stack

Python MLX Rust TypeScript Solana FastAPI GCP React

📊  Stats

 

⸸   Building sovereign AI. One weight at a time.   ⸸


HuggingFace   X   Jolyne

Popular repositories Loading

  1. turboquant-mlx turboquant-mlx Public

    First MLX implementation of TurboQuant KV cache compression for Apple Silicon

    Python 28 1

  2. RavenX-Sec RavenX-Sec Public

    Autonomous Security Intelligence Model — find, classify, remediate, and report vulnerabilities in real-time

    Python 3

  3. tesla-grok-road-agent tesla-grok-road-agent Public

    OpenClaw + Grok-4 agent that turns Tesla FSD road brainstorms into shipped code + live X posts. Ideas born on the road, posted before you park.

    JavaScript 1

  4. gemini-cli gemini-cli Public

    Forked from google-gemini/gemini-cli

    An open-source AI agent that brings the power of Gemini directly into your terminal.

    TypeScript 1

  5. triattention triattention Public

    Forked from WeianMao/triattention

    TriAttention — Efficient long reasoning with trigonometric KV cache compression. Enables OpenClaw local deployment on memory-constrained GPUs.

    Python 1

  6. ravenx-agent-safety ravenx-agent-safety Public

    Agent safety mitigations from Claude Mythos Preview system card

    Python 1