15-year-old Full-Stack Developer & Systems Engineer β shipping production-grade systems, AI tooling, and Linux automation. (and a Linux Enthusiast)
~
β― whoami
π€ Abdou (a.k.a. abdou-da0wew)
~
β― cd Me
~/Me
β― age
π’ 15 years old
~/Me
β― cd ..
~
β― cd Technical
~/Technical
β― os
π§ Linux Native
~/Technical
β― projects
π triecbot.xyz | Bots, automation & experiments
~/Technical
β― cd ..
~
β― cd Me
~/Me
β― learning
βοΈ Next.js + React + R3F (breaking stuff & fixing it)
~/Me
β― motto
β‘ Code. Learn. Improve. Repeat.
~/Me
β― cd ..
~
β― ai β Building Divee: my own AI family (LLMs, VLMs, vision, voice)
~
β― status β π’ Building & open to commissionsAI/ML: ONNX Runtime Β· embeddings Β· semantic search Β· LLM pipelines GPU: wgpu Β· GLSL Game & 3D: Unity Β· Blender
- ποΈ Systems Engineering β Backends, API gateways, streaming, circuit breakers in Go & Node
- π§ AI Engineering β Building my own AI family: LLMs, VLMs, vision & voice models
- π€ Agent Development β DiveeOI: a heavily optimized OpenCode fork with memory & dynamic resource management
- π¨ UI/UX Design β Figma, Adobe suite, clean functional interfaces
- ποΈ Database Architecture β MongoDB, Redis, PostgreSQL + object storage (S3/R2) for real-time systems
- π Security Engineering β Zero-trust design, RBAC, OAuth2, AEAD token encryption
- π± Mobile Development β Cross-platform apps with Flutter and React Native
Divee β An entire AI family, built from scratch. Not one model, not a wrapper around someone's API β an ecosystem of models and infrastructure.
- π§ LLMs β Large language models with advanced capabilities
- ποΈ VLMs β Vision-language models that see and understand
- π€ Voice Models β Speech synthesis and voice pipelines
- π· Vision Models β Image understanding and processing
- β‘ Infrastructure β The routing, serving, and tooling behind them
DiveeOI is the most featured member of the family right now: a highly optimized fork of OpenCode with my architecture on top.
- π§ Memory System β Persistent memory across sessions
- π Faster Everything β Boot speed, LLM response speed, and tool calls, all optimized
- βοΈ Dynamic Resources β Resource consumption adapts at runtime to what's needed
- π§ Improved LSP Usage β Better language server integration in the agent
Development Time: Built, broken, and rebuilt over months of real use
β Actively developed β DiveeOI evolves daily
- Triecbot β Full Discord bot with dashboards, sharding, moderation, economy
- Iris Camera β GPU-accelerated camera app (Rust, GTK4, wgpu) with face detection (Still in beta really)
- Semantic Moderation β Toxic detection from embeddings, not keywords (Rust + ONNX) (For A discord Server)
- π§ Divee family development + DiveeOI agent
- π€ Discord bot automation
- βοΈ React/Next.js projects
- π§ Linux system scripting & tooling
"In a world full of bugs, be the debugger."
Thanks for stopping by π