Skip to content
View IAmOmarJomaa's full-sized avatar
🏠
Working from home
🏠
Working from home
0 followers · 5 following

Highlights

  • Pro

Block or report IAmOmarJomaa

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 supported. This note will be visible to only you.
Report abuse

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

Report abuse
IAmOmarJomaa/README.md
github-snake

Hi, I'm Omar Jomaa. 👋

AI Practitioner | Agentic Systems & Quantitative Forecasting

I move beyond static models to build resilient, adaptive AI architectures.
Currently focused on GraphRAG, Time-Series Forecasting, and MLOps.

🚀 Engineering Highlights

🧠 Project Chimpanzee: Autonomous GraphRAG Agent

Solving the "Lost in the Middle" problem in long-context retrieval.
  • Architecture: Hybrid Retrieval (Neo4j Graph + LanceDB Vectors).
  • Key Tech: LangGraph, LangSmith, OpenAI API, Docker.
  • Result: Increased context fidelity by 15% over standard RAG baselines.

⚡ Project Vulcan: Image generation

Deploying 12B-parameter models on consumer hardware (6GB VRAM).
  • Architecture: Hybrid Pipeline (WSL2 + Windows).
  • Key Tech: 4-bit GGUF Quantization, LoRA Fine-Tuning, WebSocket Bridge.
  • Result: Reduced memory footprint by 71% (23.8GB → 6.8GB).

🛠 Tech Stack

python tensorflow pytorch pandas scikit_learn mssql aws docker c

⚡️ Connect

linkedin

Pinned Loading

  1. project_chimpanzee project_chimpanzee Public

    An adaptive GraphRAG system featuring a self-correcting LangGraph state machine and a Polars-optimized Medallion ETL pipeline for large-scale knowledge extraction from Joe Rogan Podcast.

    Python

  2. Prosthetic-AI-Core Prosthetic-AI-Core Public

    Research & Development of Proportional Myoelectric Control Systems. A systematic investigation into the accuracy-latency trade-offs between Temporal Feature-Engineered LSTMs and End-to-End CNN Spec…

    Python

  3. Project-Vulcan Project-Vulcan Public

    Hybrid Cloud-Edge orchestration pipeline for 12B-parameter Flux.1 Digital Twins. Optimized for 6GB VRAM hardware via 4-bit GGUF quantization and WSL2-to-Windows WebSocket bridging.

    Python

  4. NeuroCursor NeuroCursor Public

    A physics-based AI mouse interface. Uses MediaPipe, ONNX, and a custom kinematics engine to enable jitter free touchless control. strictly typed and production ready.

    Python