Skip to content
View youhanasheriff's full-sized avatar

Block or report youhanasheriff

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
youhanasheriff/README.md

Hi, I'm Youhana Sheriff 👋

Senior Software Engineer — Edge-AI & LLM Systems

I build AI that runs in the real world: on-device & edge inference on constrained hardware, LLM/agent/RAG systems, and the full-stack apps that ship them to users. Electronics & Communication Engineer (B.E.), 5+ years and 20+ projects across mobile, web, and cloud.

🔭 edge/on-device AI · 🤖 LLM agents & RAG · 🎙️ real-time voice & vision · 📱 full-stack delivery


🧠 AI engineering & open source

Edge & on-device AI

Project What it is
models-edge-devices Split-architecture edge AI: YOLO → LoRA-tuned TinyLlama on NPU/BPU; INT8/GGUF quantization, llama.cpp, MLX
pocket_ai Real-time YOLO11 obstacle detection for edge devices — depth estimation, spatial reasoning, spoken guidance (<30ms)

LLM systems, agents & RAG

Project What it is
SimpleMem · @sheriax/simplemem Lifelong memory for LLM agents — semantic compression + vector search, ~30× fewer tokens than full-context
llm-rag RAG over your PDFs: embeddings + vector search + LLM Q&A
llm-council Multi-agent LLM consensus: models answer, peer-review anonymously, a Chairman synthesizes

Real-time voice & media AI

Project What it is
ai-voice-call-plugin React voice-call widget on the Gemini Live API — accessible, themeable, published to npm
video-composer AI short-form video pipeline: TTS + Whisper + GPT-4 Vision → TikTok / Shorts / Reels
audio-transcription Upload/record audio → AI transcription

🚀 AI products shipped to users

  • Kizu — AI personal-finance app: ML Kit OCR receipt scanning + Claude for transaction extraction & insights (Flutter)
  • Chatiko AI — AI chatbot companion, live on Google Play
  • Talky — language-learning app with AI content moderation

📦 Published packages

  • @sheriax/simplemem (npm) — LLM agent memory + vector search
  • aws-nuke-all (npm) — delete all AWS resources across regions
  • flutter_stripe_connect (pub.dev) — Stripe Connect for Flutter
  • translations_code_gen (pub.dev) — type-safe translation codegen for Dart

🧩 Also build (full-stack & mobile)

Drawink (collaborative whiteboard) · Cubanin & Echify (social commerce) · RouteX (logistics, offline-first GPS) · Nidaa & TNTJ apps — cross-platform with Flutter, React Native (Expo), Next.js.

🧰 Tech

AI/ML: LLMs (Claude · Gemini · OpenAI · OpenRouter · local) · RAG · vector search · LoRA/MLX fine-tuning · llama.cpp · ONNX · YOLO · ML Kit Languages: TypeScript · Python · Dart · Rust · Kotlin Frameworks: Next.js · React · React Native (Expo) · Flutter · Node.js / NestJS / Fastify Cloud/Infra: AWS · GCP · Firebase · Docker · Kubernetes · Cloudflare · GitHub Actions

🌐 youhanasheriff.com · ✍️ youhanasheriff.com/blog · 💼 LinkedIn

Popular repositories Loading

  1. translations_code_gen translations_code_gen Public

    Dart package to generate translations keys and values

    Dart 6

  2. personal_expenses_app personal_expenses_app Public

    Dart 2

  3. llm-rag llm-rag Public

    TypeScript 2

  4. tntj_mosque_app tntj_mosque_app Public

    Dart 1

  5. rustgrep rustgrep Public

    Rust 1

  6. deadlock-os_scheduler deadlock-os_scheduler Public

    Jupyter Notebook 1